feat: snowpipe streaming #5110
feat: snowpipe streaming #5110
Conversation
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
8 times, most recently
from
68163a6 to
5607206
Compare
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
5 times, most recently
from
444da09 to
f569ead
Compare
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
5 times, most recently
from
788f64d to
d44d456
Compare
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
3 times, most recently
from
74e9353 to
0ccab91
Compare
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
from
0ccab91 to
23bd7f8
Compare
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
2 times, most recently
from
7a44095 to
06e0dd3
Compare
| whutils "github.com/rudderlabs/rudder-server/warehouse/utils" | ||
| ) | ||
|
|
||
| func (m *Manager) prepareChannelResponse( |
There was a problem hiding this comment.
name doesn't seems to be apt as it is doing lot of operations inside
There was a problem hiding this comment.
Renamed to initializeChannelWithSchema.
router/batchrouter/asyncdestinationmanager/snowpipestreaming/discards.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/discards.go
Outdated
Show resolved
Hide resolved
...r/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api_integration_test.go
Outdated
Show resolved
Hide resolved
...r/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api_integration_test.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/model/model.go
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/poll.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/poll.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/types.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/uploadstats.go
Outdated
Show resolved
Hide resolved
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
2 times, most recently
from
0db354f to
91341e3
Compare
|
@coderabbitai review |
WalkthroughThis pull request introduces a series of changes across multiple files, primarily enhancing the functionality and testability of the Snowpipe streaming integration within the batch router. Key modifications include the addition of new methods and structs for managing Snowpipe channels, improved error handling, and updates to existing methods for better time management and job processing. The changes also expand the set of recognized asynchronous destinations and introduce new constants and interfaces to support the Snowpipe streaming context. Overall, the updates aim to refine data handling and improve the robustness of the integration. Changes
Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media? 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
CodeRabbit Configuration File (
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 21
🧹 Outside diff range and nitpick comments (44)
warehouse/integrations/snowflake/datatype_mapper_test.go (1)
Line range hint
11-27: Consider adding edge case test scenariosWhile the current test cases cover basic scenarios, consider adding tests for:
- Maximum/minimum INT values
- Edge cases where scale affects precision
- Different VARCHAR lengths
Here's a suggested addition to the test cases:
func TestCalculateDataType(t *testing.T) { testCases := []struct { columnType string numericScale sql.NullInt64 expected string exists bool }{ {"VARCHAR", sql.NullInt64{}, "string", true}, {"INT", sql.NullInt64{}, "int", true}, {"INT", sql.NullInt64{Int64: 2, Valid: true}, "float", true}, {"UNKNOWN", sql.NullInt64{}, "", false}, + {"VARCHAR(MAX)", sql.NullInt64{}, "string", true}, + {"INT", sql.NullInt64{Int64: 10, Valid: true}, "float", true}, + {"NUMBER", sql.NullInt64{Int64: 0, Valid: true}, "int", true}, }router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/codes.go (1)
3-12: Consider adding documentation and grouping related errorsThe error constants are well-defined and follow a consistent naming pattern. However, consider these improvements for better maintainability:
- Add godoc comments for each error constant to explain when they occur
- Group related constants (e.g., authorization errors together)
- Consider adding package documentation with examples of error handling
Here's a suggested improvement:
package api +// Error codes returned by the Snowpipe Streaming API const ( + // General errors ErrUnknownError = "ERR_UNKNOWN_ERROR" ErrValidationError = "ERR_VALIDATION_ERROR" ErrSchemaConflict = "ERR_SCHEMA_CONFLICT" + // Authentication and authorization errors ErrAuthenticationFailed = "ERR_AUTHENTICATION_FAILED" ErrRoleDoesNotExistOrNotAuthorized = "ERR_ROLE_DOES_NOT_EXIST_OR_NOT_AUTHORIZED" ErrDatabaseDoesNotExistOrNotAuthorized = "ERR_DATABASE_DOES_NOT_EXIST_OR_NOT_AUTHORIZED" ErrSchemaDoesNotExistOrNotAuthorized = "ERR_SCHEMA_DOES_NOT_EXIST_OR_NOT_AUTHORIZED" ErrTableDoesNotExistOrNotAuthorized = "ERR_TABLE_DOES_NOT_EXIST_OR_NOT_AUTHORIZED" )warehouse/integrations/snowflake/datatype_mapper.go (1)
Line range hint
47-51: Consider adding validation for negative scale valuesWhile the logic for converting int to float when scale > 0 is correct, consider adding validation for negative scale values which could indicate data issues.
func CalculateDataType(columnType string, numericScale int64) (string, bool) { if datatype, ok := dataTypesMapToRudder[columnType]; ok { + if numericScale < 0 { + return "", false + } if datatype == "int" && numericScale > 0 { datatype = "float" } return datatype, true } return "", false }router/batchrouter/asyncdestinationmanager/snowpipestreaming/testhelper/testhelper.go (1)
54-70: Consider improving error handling in schema cleanupThe implementation is good with retry logic, but consider adding more context to error messages and handling specific Snowflake error codes.
Consider this enhancement:
func DropSchema(t testing.TB, db *sql.DB, namespace string) { t.Helper() - t.Log("dropping schema", namespace) + t.Logf("attempting to drop schema %q", namespace) require.Eventually(t, func() bool { _, err := db.ExecContext(context.Background(), fmt.Sprintf(`DROP SCHEMA %q CASCADE;`, namespace)) if err != nil { - t.Logf("error deleting schema %q: %v", namespace, err) + t.Logf("failed to drop schema %q: %v (will retry)", namespace, err) return false } + t.Logf("successfully dropped schema %q", namespace) return true }, time.Minute, time.Second, ) }warehouse/utils/uploader.go (1)
10-24: Consider breaking down the interface and adding documentationWhile the interface is well-structured, consider the following improvements:
- The interface seems to violate the Interface Segregation Principle by combining multiple responsibilities (schema management, file operations, configuration). Consider splitting it into smaller, focused interfaces.
- Add GoDoc comments for the interface and its methods as this appears to be a public API.
Example refactor:
+// SchemaManager handles warehouse schema operations +type SchemaManager interface { + IsWarehouseSchemaEmpty() bool + GetLocalSchema(ctx context.Context) (model.Schema, error) + UpdateLocalSchema(ctx context.Context, schema model.Schema) error + GetTableSchemaInWarehouse(tableName string) model.TableSchema + GetTableSchemaInUpload(tableName string) model.TableSchema +} + +// FileManager handles load file operations +type FileManager interface { + GetLoadFilesMetadata(ctx context.Context, options GetLoadFilesOptions) ([]LoadFile, error) + GetSampleLoadFileLocation(ctx context.Context, tableName string) (string, error) + GetSingleLoadFile(ctx context.Context, tableName string) (LoadFile, error) + GetLoadFileType() string +} + +// ConfigProvider handles warehouse-specific configurations +type ConfigProvider interface { + ShouldOnDedupUseNewRecord() bool + UseRudderStorage() bool + CanAppend() bool +} + +// Uploader combines all warehouse operations +type Uploader interface { + SchemaManager + FileManager + ConfigProvider +}router/batchrouter/asyncdestinationmanager/snowpipestreaming/apiadapter.go (2)
21-33: Add input validation in constructorConsider validating required dependencies to prevent nil pointer dereferences at runtime.
func newApiAdapter( logger logger.Logger, statsFactory stats.Stats, api api, destination *backendconfig.DestinationT, ) api { + if logger == nil { + panic("logger is required") + } + if statsFactory == nil { + panic("statsFactory is required") + } + if api == nil { + panic("api is required") + } + if destination == nil { + panic("destination is required") + } return &apiAdapter{ logger: logger, statsFactory: statsFactory, api: api, destination: destination, } }
122-124: Define metric name as a constantConsider moving the metric name to a constant to maintain consistency and ease future updates.
+const ( + snowpipeStreamingAPIResponseTimeMetric = "snowpipe_streaming_api_response_time" +) func (a *apiAdapter) recordDuration(tags stats.Tags) func() { - return a.statsFactory.NewTaggedStat("snowpipe_streaming_api_response_time", stats.TimerType, tags).RecordDuration() + return a.statsFactory.NewTaggedStat(snowpipeStreamingAPIResponseTimeMetric, stats.TimerType, tags).RecordDuration() }router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/model/model_test.go (2)
12-79: Consider adding error handling test casesThe test cases are well-structured and cover various scenarios, but consider adding test cases for:
- Invalid scale values (negative or extremely large)
- Malformed type strings
- Case sensitivity in type names
Example additional test case:
{ name: "Invalid scale value", tableSchema: map[string]ColumnInfo{ "column1": {Type: lo.ToPtr("NUMBER(2,0)"), Scale: lo.ToPtr(-1.0)}, }, expected: whutils.ModelTableSchema{}, },
1-125: Add package and function documentationConsider adding:
- Package documentation explaining the purpose of the model package
- Function documentation for both test functions describing what aspects they're testing
Example documentation:
// Package model provides types and functions for handling Snowpipe streaming operations // TestChannelResponse_GenerateSnowPipeSchema validates the schema generation logic // for various Snowflake data types and edge cases // TestChannelResponse_UnmarshalJSON ensures proper parsing of Snowpipe channel // responses including both successful and error scenariosrouter/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming_test.go (1)
12-113: Well-structured test cases with good coverage.The table-driven tests effectively cover the main scenarios for column comparison, including type differences, case sensitivity, and edge cases with empty/nil schemas.
Consider adding edge cases for column names.
While the current test cases are comprehensive for typical scenarios, consider adding tests for:
- Column names with special characters
- Very long column names
- Unicode characters in column names
{ + name: "column names with special characters and unicode", + eventSchema: whutils.ModelTableSchema{ + "special@#$": "STRING", + "very_long_column_name_exceeding_typical_length": "STRING", + "column_名字": "STRING", + }, + snowPipeSchema: whutils.ModelTableSchema{}, + expected: []whutils.ColumnInfo{ + {Name: "special@#$", Type: "STRING"}, + {Name: "very_long_column_name_exceeding_typical_length", Type: "STRING"}, + {Name: "column_名字", Type: "STRING"}, + }, + },router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api_integration_test.go (3)
49-60: Consider adding documentation for required environment variables.While the code correctly checks for required environment variables, it would be helpful to add a comment explaining what
TestKeyPairUnencryptedrepresents and its expected format.
61-156: Consider extracting test data and configuration to test fixtures.The test is well-structured but could benefit from some improvements:
- Move the test data (rows) to a separate test fixture file
- Extract configuration values (ports, timeouts) to named constants
- Consider using a test helper for common setup steps (creating namespace and table)
Example for constants:
const ( snowpipeClientPort = 9078 statusCheckTimeout = 30 * time.Second statusCheckInterval = 300 * time.Millisecond )
227-250: Add documentation for helper functions.While the functions are well-implemented, they would benefit from documentation explaining their purpose and parameters:
// prepareDestination creates a new destination configuration for Snowpipe streaming // using the provided credentials. func prepareDestination(credentials *testhelper.TestCredentials) backendconfig.DestinationT { // ... existing implementation ... } // convertRowsToRecord converts a slice of model.Row to a slice of string slices, // maintaining the order specified by the columns parameter. Missing values are // represented as empty strings. func convertRowsToRecord(rows []model.Row, columns []string) [][]string { // ... existing implementation ... }warehouse/slave/worker.go (1)
Line range hint
525-573: Consider improving error handling and type safetyThe function could benefit from several improvements:
- Add more descriptive error messages including the actual value that failed conversion
- Add panic recovery for type assertions
- Consider breaking down the function into smaller, more focused functions for each type conversion
Here's a suggested improvement:
func HandleSchemaChange(existingDataType, currentDataType model.SchemaType, value any) (any, error) { + // Recover from any panics during type assertions + defer func() { + if r := recover(); r != nil { + err = fmt.Errorf("panic during type conversion: %v", r) + } + }() + var ( newColumnVal any err error ) if existingDataType == model.StringDataType || existingDataType == model.TextDataType { // only stringify if the previous type is non-string/text/json if currentDataType != model.StringDataType && currentDataType != model.TextDataType && currentDataType != model.JSONDataType { newColumnVal = fmt.Sprintf("%v", value) } else { newColumnVal = value } } else if (currentDataType == model.IntDataType || currentDataType == model.BigIntDataType) && existingDataType == model.FloatDataType { intVal, ok := value.(int) if !ok { - err = fmt.Errorf("incompatible schema conversion from %v to %v", existingDataType, currentDataType) + err = fmt.Errorf("incompatible schema conversion from %v to %v (value: %v of type %T)", + existingDataType, currentDataType, value, value) } else { newColumnVal = float64(intVal) } } else if currentDataType == model.FloatDataType && (existingDataType == model.IntDataType || existingDataType == model.BigIntDataType) { floatVal, ok := value.(float64) if !ok { - err = fmt.Errorf("incompatible schema conversion from %v to %v", existingDataType, currentDataType) + err = fmt.Errorf("incompatible schema conversion from %v to %v (value: %v of type %T)", + existingDataType, currentDataType, value, value) } else { newColumnVal = int(floatVal) } } else if existingDataType == model.JSONDataType { var interfaceSliceSample []any if currentDataType == model.IntDataType || currentDataType == model.FloatDataType || currentDataType == model.BooleanDataType { newColumnVal = fmt.Sprintf("%v", value) } else if reflect.TypeOf(value) == reflect.TypeOf(interfaceSliceSample) { newColumnVal = value } else { newColumnVal = fmt.Sprintf(`"%v"`, value) } } else { - err = fmt.Errorf("incompatible schema conversion from %v to %v", existingDataType, currentDataType) + err = fmt.Errorf("incompatible schema conversion from %v to %v (value: %v of type %T)", + existingDataType, currentDataType, value, value) } return newColumnVal, err }Additionally, consider extracting type-specific conversions into separate functions:
func handleStringConversion(value any) any { return fmt.Sprintf("%v", value) } func handleFloatToIntConversion(value any) (any, error) { floatVal, ok := value.(float64) if !ok { return nil, fmt.Errorf("value %v is not a float64", value) } return int(floatVal), nil }gateway/handle_lifecycle.go (4)
Line range hint
269-278: Consider adding timeout to errgroup contextThe
errgroupis created without a timeout, which could lead to goroutines running indefinitely if the parent context isn't cancelled. Consider adding a timeout to the errgroup context.-g, _ := errgroup.WithContext(ctx) +g, gctx := errgroup.WithContext(ctx) +gctx, cancel := context.WithTimeout(gctx, 30*time.Second) +defer cancel()
Line range hint
449-452: Enhance error handling in Shutdown methodThe
Shutdownmethod should handle errors frombackgroundWait()more gracefully and potentially implement a timeout to prevent hanging.func (gw *Handle) Shutdown() error { gw.backgroundCancel() + shutdownCtx, cancel := context.WithTimeout(context.Background(), 30*time.Second) + defer cancel() + + errChan := make(chan error, 1) + go func() { + errChan <- gw.webhook.Shutdown() + }() + + select { + case err := <-errChan: + if err != nil { + return fmt.Errorf("webhook shutdown failed: %w", err) + } + case <-shutdownCtx.Done(): + return fmt.Errorf("shutdown timed out") + } + + gw.inFlightRequests.Wait() + + for _, worker := range gw.userWebRequestWorkers { + close(worker.webRequestQ) + } + + if err := gw.backgroundWait(); err != nil { + return fmt.Errorf("background tasks shutdown failed: %w", err) + } + return nil - if err := gw.webhook.Shutdown(); err != nil { - return err - } - - gw.inFlightRequests.Wait() - - // UserWebRequestWorkers - for _, worker := range gw.userWebRequestWorkers { - close(worker.webRequestQ) - } - - return gw.backgroundWait() }
Line range hint
308-321: Consider implementing rate limiting for batch operationsThe batch processing logic in
userWorkerRequestBatchercould benefit from rate limiting to prevent overwhelming the system during high load scenarios.Consider implementing a token bucket rate limiter or similar mechanism to control the rate of batch processing. This would help prevent resource exhaustion during traffic spikes.
Line range hint
366-368: Add metrics for job processing errorsThe error handling in
dbWriterWorkerProcessshould include metrics to track job processing failures for better observability.if err != nil { + gw.stats.NewStat("gateway.db_writer_errors", stats.CountType).Count(1) errorMessage := err.Error() if ctx.Err() != nil { errorMessage = ctx.Err().Error() }processor/transformer/transformer.go (1)
540-542: Consider refactoring URL parameter construction to reduce duplication.The URL parameter construction for warehouse destinations, including Snowpipe Streaming, follows a similar pattern. Consider extracting the common logic into a helper function to improve maintainability.
Here's a suggested refactor:
+func (trans *handle) getWarehouseQueryParams() string { + return fmt.Sprintf("whSchemaVersion=%s&whIDResolve=%v", + trans.conf.GetString("Warehouse.schemaVersion", "v1"), + warehouseutils.IDResolutionEnabled()) +} func (trans *handle) destTransformURL(destType string) string { destinationEndPoint := fmt.Sprintf("%s/v0/destinations/%s", trans.config.destTransformationURL, strings.ToLower(destType)) if _, ok := warehouseutils.WarehouseDestinationMap[destType]; ok { - whSchemaVersionQueryParam := fmt.Sprintf("whSchemaVersion=%s&whIDResolve=%v", trans.conf.GetString("Warehouse.schemaVersion", "v1"), warehouseutils.IDResolutionEnabled()) + whSchemaVersionQueryParam := trans.getWarehouseQueryParams() switch destType { case warehouseutils.RS: return destinationEndPoint + "?" + whSchemaVersionQueryParam case warehouseutils.CLICKHOUSE: enableArraySupport := fmt.Sprintf("chEnableArraySupport=%s", fmt.Sprintf("%v", trans.conf.GetBool("Warehouse.clickhouse.enableArraySupport", false))) return destinationEndPoint + "?" + whSchemaVersionQueryParam + "&" + enableArraySupport default: return destinationEndPoint + "?" + whSchemaVersionQueryParam } } if destType == warehouseutils.SnowpipeStreaming { - return fmt.Sprintf("%s?whSchemaVersion=%s&whIDResolve=%t", destinationEndPoint, trans.conf.GetString("Warehouse.schemaVersion", "v1"), warehouseutils.IDResolutionEnabled()) + return destinationEndPoint + "?" + trans.getWarehouseQueryParams() } return destinationEndPoint }warehouse/utils/utils.go (1)
222-225: Add documentation for type aliasesWhile the type aliases are well-named, adding documentation comments would improve clarity by explaining the purpose of these aliases and when to use them over the original types.
type ( + // ModelWarehouse is an alias for model.Warehouse to simplify type references ModelWarehouse = model.Warehouse + // ModelTableSchema is an alias for model.TableSchema to simplify type references ModelTableSchema = model.TableSchema )gateway/handle.go (1)
491-491: Consider refactoring for better maintainability.The
getJobDataFromRequestfunction handles multiple responsibilities and is quite long. Consider breaking it down into smaller, focused functions:
- Event validation
- Batch processing
- Error handling
- User suppression checks
Example refactoring approach:
func (gw *Handle) getJobDataFromRequest(req *webRequestT) (jobData *jobFromReq, err error) { + if err := gw.validateRequest(req); err != nil { + return nil, err + } + + jobData, err = gw.processEventBatch(req) + if err != nil { + return nil, err + } + + if err := gw.applyRateLimiting(req, jobData); err != nil { + return nil, err + } + + return jobData, nil } +func (gw *Handle) validateRequest(req *webRequestT) error { + // Extract request validation logic +} + +func (gw *Handle) processEventBatch(req *webRequestT) (*jobFromReq, error) { + // Extract batch processing logic +} + +func (gw *Handle) applyRateLimiting(req *webRequestT, jobData *jobFromReq) error { + // Extract rate limiting logic +}warehouse/utils/reservedkeywords.go (1)
97-189: Consider reducing code duplication with SNOWFLAKE keywords.Since SNOWPIPE_STREAMING uses identical keywords to SNOWFLAKE, consider refactoring to reuse the same keyword map. This would reduce maintenance overhead and prevent potential divergence.
Example refactor:
var snowflakeKeywords = map[string]bool{ "ACCOUNT": true, "ALL": true, // ... rest of the keywords } var ReservedKeywords = map[string]map[string]bool{ "SNOWFLAKE": snowflakeKeywords, - "SNOWPIPE_STREAMING": { - "ACCOUNT": true, - "ALL": true, - // ... duplicate keywords - }, + "SNOWPIPE_STREAMING": snowflakeKeywords, // ... rest of the destinations }router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/model/model.go (3)
93-95: Wrap unmarshalling errors with context.Consider wrapping the error returned by
json.Unmarshalto provide additional context for debugging.Apply this diff:
if err := json.Unmarshal(data, &temp); err != nil { - return err + return fmt.Errorf("failed to unmarshal ChannelResponse: %w", err) }Also, add the necessary import:
import ( "strings" "sync" "fmt" // Added import jsoniter "github.com/json-iterator/go" "github.com/rudderlabs/rudder-server/warehouse/integrations/snowflake" whutils "github.com/rudderlabs/rudder-server/warehouse/utils" )
81-83: Re-evaluate the use of a package-level variable fordataTypeCache.Consider encapsulating
dataTypeCachewithin a struct or providing access methods to improve modularity and prevent unintended access.
126-134: Handle type assertion inextractDataTypesafely.Although the type assertion
cachedResult.(string)is expected to be safe, consider adding a type check to prevent potential panics.Apply this diff:
if cachedResult, found := dataTypeCache.Load(input); found { - return cachedResult.(string) + if resultStr, ok := cachedResult.(string); ok { + return resultStr + } }router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api.go (2)
42-70: Refactor to reduce code duplication in HTTP request handlingThe methods
CreateChannel,DeleteChannel,GetChannel,Insert, andGetStatusshare similar patterns for creating HTTP requests, setting headers, sending requests, and handling responses. This code duplication can be minimized by abstracting common logic into helper functions.Consider creating helper functions for:
- Building request URLs: A function to construct URLs based on endpoint paths and parameters.
- Creating HTTP requests: A helper to create requests with common headers and context.
- Executing requests and handling responses: A generic function to execute the request, check status codes, read the response body, and handle errors.
- Decoding JSON responses: A utility to decode JSON responses into the appropriate structs.
This refactoring will improve maintainability and reduce the potential for inconsistencies or errors.
Also applies to: 72-96, 98-121, 123-151, 153-176
78-78: Remove unnecessaryContent-Typeheader from requests without a bodySetting the
Content-Typeheader on HTTP requests that do not include a body (such asGETandDELETErequests without a payload) is unnecessary and can be omitted.Consider removing the
Content-Typeheader from these requests to simplify them:-req.Header.Set("Content-Type", "application/json")This change clarifies that these requests do not send any content in the body.
Also applies to: 104-104, 159-159
router/batchrouter/asyncdestinationmanager/snowpipestreaming/channel.go (2)
65-67: Simplify defer statements by callingCleanupdirectlyYou can simplify your defer statements by directly deferring the
Cleanupcall without wrapping it in an anonymous function:- defer func() { - snowflakeManager.Cleanup(ctx) - }() + defer snowflakeManager.Cleanup(ctx)This improves code readability and maintains the same functionality.
Also applies to: 158-160, 186-188
142-168: Refactor duplicated code inhandleSchemaErrorandhandleTableErrorThe functions
handleSchemaErrorandhandleTableErrorshare similar logic for creating the Snowflake manager, deferring cleanup, and creating the schema or table. Consider refactoring the common code into a helper function to reduce duplication and enhance maintainability.You could create a helper function like this:
func (m *Manager) handleError( ctx context.Context, channelReq *model.CreateChannelRequest, eventSchema whutils.ModelTableSchema, createSchema bool, ) (*model.ChannelResponse, error) { snowflakeManager, err := m.createSnowflakeManager(ctx, channelReq.TableConfig.Schema) if err != nil { return nil, fmt.Errorf("creating snowflake manager: %w", err) } defer snowflakeManager.Cleanup(ctx) if createSchema { if err := snowflakeManager.CreateSchema(ctx); err != nil { return nil, fmt.Errorf("creating schema: %w", err) } } if err := snowflakeManager.CreateTable(ctx, channelReq.TableConfig.Table, eventSchema); err != nil { return nil, fmt.Errorf("creating table: %w", err) } return m.api.CreateChannel(ctx, channelReq) }And then modify
handleSchemaErrorandhandleTableErrorto use this helper:func (m *Manager) handleSchemaError( ctx context.Context, channelReq *model.CreateChannelRequest, eventSchema whutils.ModelTableSchema, ) (*model.ChannelResponse, error) { m.logger.Infon("Handling schema error", logger.NewStringField("schema", channelReq.TableConfig.Schema), logger.NewStringField("table", channelReq.TableConfig.Table), ) return m.handleError(ctx, channelReq, eventSchema, true) } func (m *Manager) handleTableError( ctx context.Context, channelReq *model.CreateChannelRequest, eventSchema whutils.ModelTableSchema, ) (*model.ChannelResponse, error) { m.logger.Infon("Handling table error", logger.NewStringField("schema", channelReq.TableConfig.Schema), logger.NewStringField("table", channelReq.TableConfig.Table), ) return m.handleError(ctx, channelReq, eventSchema, false) }Also applies to: 170-193
router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api_test.go (4)
359-359: Fix typo in test name: "Get Statu" should be "Get Status"There's a typo in the test name. It should be "Get Status" to correctly reflect the method being tested.
41-84: Refactor repeatedhttptest.NewServersetups into helper functionsThe setup code for
httptest.NewServerwith similar handler functions is repeated across multiple test cases (Create Channel,Delete Channel,Get Channel,Insert,Get Status). Refactoring this code into reusable helper functions would reduce code duplication and improve maintainability.
55-56: SetContent-Typeheader in HTTP responsesIn the test HTTP handlers, it's good practice to set the
Content-Typeheader when sending a JSON response to simulate real-world API behavior. Please addw.Header().Set("Content-Type", "application/json")before writing the response. This applies to all handlers returning JSON responses.
127-127: Correct the test case descriptionThe test case description should be "Request failure (non-200 status code)" without the apostrophe. This change improves clarity and grammatical correctness.
router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming.go (6)
13-13: Nitpick: Unnecessary Import ofstrconvThe
strconvpackage is imported on line 12~ but does not appear to be used anywhere in the code.Consider removing the unused import to keep the code clean:
- "strconv"
287-289: Error Handling Improvement: Clarify Error Message When Creating ChannelIn the error message returned when failing to create a channel, consider including the
destinationIDor more context to aid in debugging.Update the error message as follows:
- return nil, nil, fmt.Errorf("creating channel %s: %w", info.tableName, err) + return nil, nil, fmt.Errorf("creating channel for table %s in destination %s: %w", info.tableName, destinationID, err)
464-465: Error Handling Improvement: More Informative Error on Invalid Status ResponseWhen the status response is invalid or unsuccessful, returning a generic
errInvalidStatusResponsemay not provide enough context.Enhance the error message to include details from
statusRes:if !statusRes.Valid || !statusRes.Success { - return false, errInvalidStatusResponse + return false, fmt.Errorf("invalid status response: success=%v, valid=%v", statusRes.Success, statusRes.Valid) }
319-319: Error Message Enhancement: Include Error DetailsIn the error message constructed when inserting data fails, consider including the error messages from
insertRes.Errorsfor better troubleshooting.Modify the error message to include detailed errors:
return nil, nil, fmt.Errorf( - "inserting data %s failed: %s %v", + "inserting data for table %s failed: code=%s, errors=%v", info.tableName, insertRes.Code, errorMessages, )
12-12: Unused Import:slicesPackage Not UtilizedThe
slicespackage is imported but not used in the code.Consider removing the unused import to clean up dependencies:
- "slices"
181-184: Variable Shadowing: Overlapping Variable NamesIn lines 181-184, the variables
importingJobIDsandfailedJobIDsare declared, which might shadow any variables with the same names in the broader scope.Ensure that variable names are unique within their scope to prevent shadowing and potential bugs.
router/batchrouter/handle_async.go (1)
Line range hint
323-348: Replacetime.Afterwithtime.Tickerto prevent timer leaksUsing
time.After()inside a loop can lead to unbounded accumulation of timers, causing memory leaks, because each call creates a new timer that cannot be garbage collected until it fires. It is recommended to usetime.Tickerfor periodic actions within loops.Apply this diff to refactor the code:
func (brt *Handle) asyncUploadWorker(ctx context.Context) { if !asynccommon.IsAsyncDestination(brt.destType) { return } + ticker := time.NewTicker(brt.asyncUploadWorkerTimeout.Load()) + defer ticker.Stop() + for { select { case <-ctx.Done(): return - case <-time.After(brt.asyncUploadWorkerTimeout.Load()): + case <-ticker.C: brt.configSubscriberMu.RLock() destinationsMap := brt.destinationsMap uploadIntervalMap := brt.uploadIntervalMap brt.configSubscriberMu.RUnlock() for destinationID := range destinationsMap { _, ok := brt.asyncDestinationStruct[destinationID] if !ok || brt.asyncDestinationStruct[destinationID].UploadInProgress { continue } timeElapsed := time.Since(brt.asyncDestinationStruct[destinationID].CreatedAt) brt.asyncDestinationStruct[destinationID].UploadMutex.Lock() timeout := uploadIntervalMap[destinationID] if brt.asyncDestinationStruct[destinationID].Exists && (brt.asyncDestinationStruct[destinationID].CanUpload || timeElapsed > timeout) { brt.asyncDestinationStruct[destinationID].CanUpload = true brt.asyncDestinationStruct[destinationID].PartFileNumber++ uploadResponse := brt.asyncDestinationStruct[destinationID].Manager.Upload(brt.asyncDestinationStruct[destinationID]) if uploadResponse.ImportingParameters != nil && len(uploadResponse.ImportingJobIDs) > 0 { brt.asyncDestinationStruct[destinationID].UploadInProgress = true } brt.setMultipleJobStatus(uploadResponse, true, brt.asyncDestinationStruct[destinationID].AttemptNums, brt.asyncDestinationStruct[destinationID].FirstAttemptedAts, brt.asyncDestinationStruct[destinationID].OriginalJobParameters) brt.asyncStructCleanUp(destinationID) } brt.asyncDestinationStruct[destinationID].UploadMutex.Unlock() } } } }router/batchrouter/handle.go (4)
81-81: Add documentation forasyncUploadWorkerTimeout.The new field
asyncUploadWorkerTimeout config.ValueLoader[time.Duration]has been added to theHandlestruct. Adding a comment to explain its purpose, default value, and how it influences the async upload worker would enhance code readability and maintainability.
154-154: Initializenowfunction in a thread-safe manner.The
nowfunction is initialized in themainLoopmethod withbrt.now = timeutil.Nowif it isnil. Ensure that this lazy initialization is thread-safe. Alternatively, consider initializingbrt.nowduring the construction of theHandlestruct to avoid the nil check in the main loop.
Line range hint
422-436: Replacepaniccalls with proper error handling.In the
uploadmethod, several error conditions are handled usingpanic(err). Usingpanicis discouraged in production code as it can crash the entire application. Instead, handle the errors gracefully by returning them to the caller or logging them appropriately.Apply this diff to handle errors without panicking:
err = os.MkdirAll(filepath.Dir(gzipFilePath), os.ModePerm) if err != nil { - panic(err) + brt.logger.Errorf("Failed to create directories for %s: %v", gzipFilePath, err) + return UploadResult{ + Error: err, + LocalFilePaths: []string{gzipFilePath}, + } } err = misc.CreateGZ(gzipFilePath) if err != nil { - panic(err) + brt.logger.Errorf("Failed to create gzip writer for %s: %v", gzipFilePath, err) + return UploadResult{ + Error: err, + LocalFilePaths: []string{gzipFilePath}, + } }Ensure all instances of
panic(err)are replaced with appropriate error handling throughout the method.
Line range hint
655-655: Fix typo in comment forskipFetchingJobsmethod.There is a typo in the comment:
-// skipFetchingJobs returns true if the destination type is async and the there are still jobs in [importing] state for this destination type +// skipFetchingJobs returns true if the destination type is async and there are still jobs in [importing] state for this destination type
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
⛔ Files ignored due to path filters (3)
.github/workflows/tests.yamlis excluded by!**/*.yamlintegration_test/snowpipestreaming/testdata/docker-compose.rudder-transformer.ymlis excluded by!**/*.ymlrouter/batchrouter/asyncdestinationmanager/snowpipestreaming/testdata/docker-compose.rudder-snowpipe-clients.ymlis excluded by!**/*.yml
📒 Files selected for processing (31)
gateway/handle.go(1 hunks)gateway/handle_lifecycle.go(3 hunks)processor/transformer/transformer.go(1 hunks)router/batchrouter/asyncdestinationmanager/common/utils.go(1 hunks)router/batchrouter/asyncdestinationmanager/manager.go(2 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/apiadapter.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/channel.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/discards.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api_integration_test.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api_test.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/codes.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/model/model.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/model/model_test.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming_test.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/testhelper/testhelper.go(1 hunks)router/batchrouter/asyncdestinationmanager/snowpipestreaming/types.go(1 hunks)router/batchrouter/handle.go(4 hunks)router/batchrouter/handle_async.go(3 hunks)router/batchrouter/handle_lifecycle.go(2 hunks)utils/misc/misc.go(1 hunks)warehouse/integrations/manager/manager.go(2 hunks)warehouse/integrations/snowflake/datatype_mapper.go(1 hunks)warehouse/integrations/snowflake/datatype_mapper_test.go(1 hunks)warehouse/integrations/snowflake/snowflake.go(1 hunks)warehouse/slave/worker.go(2 hunks)warehouse/slave/worker_test.go(1 hunks)warehouse/utils/reservedkeywords.go(1 hunks)warehouse/utils/uploader.go(1 hunks)warehouse/utils/utils.go(3 hunks)
🔇 Additional comments (47)
router/batchrouter/asyncdestinationmanager/common/utils.go (1)
6-6: LGTM! Verify constant usage consistency.
The addition of "SNOWPIPE_STREAMING" to asyncDestinations is appropriate for the new feature. However, let's verify the constant usage across the codebase.
✅ Verification successful
LGTM! The constant usage is consistent across the codebase.
The verification shows that "SNOWPIPE_STREAMING" is consistently used across the codebase:
- Defined as a constant in
warehouse/utils/utils.goasSnowpipeStreaming = "SNOWPIPE_STREAMING" - Used consistently in batch destinations list in
utils/misc/misc.go - Referenced correctly in all integration points including warehouse manager, transformer, and async destination manager
- Properly used in test files with the exact same string value
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify consistent usage of SNOWPIPE_STREAMING constant
# Expected: All occurrences should use the exact same string "SNOWPIPE_STREAMING"
# Search for any variations of the constant name to ensure consistency
rg -i "snowpipe.*streaming" --type go
# Search for destination type declarations to ensure alignment
ast-grep --pattern 'const $_ = "SNOWPIPE_STREAMING"'
Length of output: 40376
warehouse/integrations/snowflake/datatype_mapper_test.go (1)
24-24: Verify null handling safety
The change from passing tc.numericScale to tc.numericScale.Int64 moves null checking responsibility to the caller. Ensure that CalculateDataType properly handles zero values when numericScale.Valid is false.
✅ Verification successful
Let me gather more information about the implementation to verify null handling safety.
Let me fix the regex and try again to check the implementation.
Accessing numericScale.Int64 is safe in this context
The implementation is safe because:
CalculateDataTypeonly usesnumericScaleto convert "int" to "float" when scale > 0- When
numericScale.Validis false,Int64defaults to 0, which maintains the correct behavior - Test cases cover both scenarios:
sql.NullInt64{}(zero value, Valid=false) for VARCHAR and INT typessql.NullInt64{Int64: 2, Valid: true}for INT with scale
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Check if CalculateDataType is called with invalid sql.NullInt64 values in the codebase
# Search for all calls to CalculateDataType
ast-grep --pattern 'CalculateDataType($_, $_)'
# Search for null checks before calls
rg -B 5 'CalculateDataType\('
Length of output: 2992
Script:
#!/bin/bash
# Check the implementation of CalculateDataType and how it handles numericScale
ast-grep --pattern 'func CalculateDataType($_, $_) {
$$$
}'
# Check if numericScale is used in any conditional logic
rg -A 10 'numericScale.*{' warehouse/integrations/snowflake/
Length of output: 220
Script:
#!/bin/bash
# Check the implementation of CalculateDataType and its handling of numericScale
cat warehouse/integrations/snowflake/datatype_mapper.go
# Check all usages of numericScale in conditional statements
rg -A 5 'numericScale' warehouse/integrations/snowflake/
Length of output: 4346
router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/codes.go (2)
1-2: LGTM! Package structure follows Go conventions
The package is appropriately named and correctly placed under the internal directory, indicating it's meant for internal use only.
3-12: Verify coverage of all Snowpipe error scenarios
Let's ensure all potential Snowpipe streaming error scenarios are covered by these error codes.
✅ Verification successful
Let me analyze the error handling patterns in the codebase to verify if all necessary error scenarios are covered by the error codes.
Based on the analysis of the codebase, I can now provide a comprehensive verification of the error codes. Let me generate the final response.
Error codes are comprehensive and properly mapped
The current error codes in codes.go cover all critical Snowpipe streaming error scenarios:
- Authentication failures:
ErrAuthenticationFailed - Authorization/access issues:
ErrRoleDoesNotExistOrNotAuthorizedErrDatabaseDoesNotExistOrNotAuthorizedErrSchemaDoesNotExistOrNotAuthorizedErrTableDoesNotExistOrNotAuthorized
- Schema validation:
ErrValidationErrorErrSchemaConflict
- Generic errors:
ErrUnknownError
The error codes align with:
- Snowflake's API error responses (seen in API tests)
- Real-world error scenarios (validated through integration tests)
- Standard error handling patterns in the codebase
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Search for any additional error handling patterns in the codebase
# that might indicate missing error scenarios
# Look for error handling in Snowpipe related files
rg -g '**/*snow*.go' -B 2 -A 2 'error|err|fail|invalid'
# Look for TODO/FIXME comments related to error handling
rg -g '**/*snow*.go' 'TODO|FIXME|XXX' -C 2
Length of output: 134196
Script:
#!/bin/bash
# Search for error handling patterns in the Snowpipe API code
ast-grep --pattern 'const (
$$$
Err$_ = $_
$$$
)'
# Look for error handling in Snowpipe API related files
rg -g '**/snowpipestreaming/internal/api/*.go' -B 2 -A 2 'error|err|fail|invalid'
Length of output: 57187
warehouse/integrations/snowflake/datatype_mapper.go (1)
45-45: Verify null handling in callers after signature change
The function has been exported and the parameter type changed from sql.NullInt64 to int64. While this simplifies the function, ensure all callers properly handle NULL values before passing the scale.
✅ Verification successful
Let me check how numericScale is handled before being passed to CalculateDataType.
Let me check the complete context of the numericScale handling in the Snowpipe streaming model file to ensure proper null handling.
Function signature change is safe and properly handled
The change from sql.NullInt64 to int64 is well-handled across all callers:
- In
snowflake.go, the caller safely accesses.Int64from asql.NullInt64 - In
model.go, proper initialization with default0and safe pointer dereferencing is implemented - Test cases cover both valid and invalid scenarios
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Search for all callers of CalculateDataType to verify NULL handling
rg -A 5 "CalculateDataType\("
Length of output: 2729
Script:
#!/bin/bash
# Check how numericScale is handled in the calling contexts
rg -B 5 "CalculateDataType\("
Length of output: 2677
Script:
#!/bin/bash
# Check the complete context of numericScale handling in the model file
rg -B 10 -A 5 "numericScale :=" router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/model/model.go
Length of output: 432
router/batchrouter/asyncdestinationmanager/snowpipestreaming/testhelper/testhelper.go (5)
1-17: LGTM! Clean package structure and imports
The package structure follows Go conventions, and imports are well-organized with a clear separation between standard library, external, and internal packages.
19-22: LGTM! Well-defined constants
Constants are clearly named and follow the standard naming conventions for environment variables.
34-46: LGTM! Well-implemented credentials retrieval
The function has proper error handling and clear error messages. The use of os.LookupEnv is appropriate for test credentials.
48-52: LGTM! Robust schema name generation
Good implementation with:
- Secure random string generation
- Timestamp inclusion for uniqueness
- Proper case conversion for Snowflake compatibility
24-32: Consider secure handling of sensitive credentials
The struct contains sensitive information (private key and passphrase). Ensure that these credentials are properly secured in the testing environment and not accidentally logged or exposed.
Let's verify the usage of these sensitive fields:
Consider implementing the following security measures:
- Add documentation comments about the security implications
- Consider implementing secure cleanup after tests
✅ Verification successful
Based on the search results, I can generate the final response as I have enough context about how these credentials are handled in the codebase.
Sensitive credentials are properly handled in test environment
The codebase demonstrates secure handling of sensitive credentials in test environments:
- Test credentials are sourced from environment variables (not hardcoded)
- Private keys are stored in temporary directories during tests using
t.TempDir() - Private keys are properly formatted and escaped when used in configurations
- Test credentials are scoped to test functions and cleaned up after tests complete
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Search for potential logging or exposure of sensitive credentials
rg -i "privatekey|passphrase" -A 5 -B 5
Length of output: 70435
warehouse/utils/uploader.go (2)
1-9: LGTM: Clean package structure and imports
The package name and imports are appropriate for the functionality being implemented.
1-55: Verify the migration from old Uploader interface
According to the AI summary, this interface replaces a previous Uploader interface from warehouse/utils/utils.go. Let's verify the migration is complete.
✅ Verification successful
Based on the search results, I can now provide a final response about the migration of the Uploader interface:
Migration to new Uploader interface is complete
The codebase search shows that:
- The new Uploader interface in
warehouse/utils/uploader.gois being used extensively across the warehouse integrations (Snowflake, Redshift, PostgreSQL, etc.) - No references to the old interface location (
warehouse/utils/utils.go) were found - All implementations (noopUploader, dummyUploader, source.Uploader) use the new interface
- Mock implementations in tests are generated from the new interface location
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Check for any remaining references to the old Uploader interface
# Search for old interface usage
rg -l "warehouse/utils/utils" | while read -r file; do
echo "Checking $file for old Uploader interface usage..."
rg "Uploader" "$file"
done
# Verify all implementations of the new interface
ast-grep --pattern 'type $_ struct {
$$$
}
func ($_ $_) IsWarehouseSchemaEmpty() bool {
$$$
}'
Length of output: 146
Script:
#!/bin/bash
# Let's try a different approach to verify the interface implementations and usage
# Search for any implementations of the Uploader interface
echo "=== Searching for Uploader interface implementations ==="
rg "type.*struct.*{" -A 5 | rg -B 5 "IsWarehouseSchemaEmpty\(\).*bool"
# Search for imports of the old utils package
echo -e "\n=== Searching for old utils package imports ==="
rg "\"github.com/rudderlabs/rudder-server/warehouse/utils\""
# Search for Uploader type usage
echo -e "\n=== Searching for Uploader type usage ==="
rg "Uploader" -A 2 -B 2
# Check if there are any references to the old location
echo -e "\n=== Checking references to potential old location ==="
fd "utils.go" --type f | while read -r file; do
echo "Checking $file..."
rg "type.*Uploader.*interface" "$file"
done
Length of output: 110347
router/batchrouter/asyncdestinationmanager/manager.go (2)
19-19: LGTM! Import statement is correctly placed and follows conventions.
The snowpipestreaming package import is properly grouped with other internal destination manager imports.
45-46:
Review error handling and parameter consistency
A few concerns with the new case implementation:
-
Error Handling: Other destination managers (e.g., BINGADS_AUDIENCE) return potential errors from their New() function, but this implementation always returns nil error. Consider updating the snowpipestreaming.New() signature to return an error if appropriate.
-
Parameter Consistency: Other destination managers (e.g., BINGADS_AUDIENCE) receive backendConfig, but it's not passed to snowpipestreaming.New(). Verify if this parameter is needed.
Let's verify the integration:
router/batchrouter/asyncdestinationmanager/snowpipestreaming/apiadapter.go (2)
14-19: LGTM! Constants are well-defined
The API operation constants follow a consistent naming convention and are used appropriately for metrics tagging.
24-24: Verify the api interface definition
The api type used here isn't immediately visible. Let's verify its definition and contract.
✅ Verification successful
Based on the code provided in the context, I can now generate the final response as I can see the complete definition of both the api interface and its implementation.
The api interface is well-defined and properly implemented
The api interface is defined within the same package and is implemented by the apiAdapter struct. The interface defines methods for managing Snowpipe streaming channels:
CreateChannel: Creates a new channel for streaming dataDeleteChannel: Deletes an existing channelInsert: Inserts data into a channelGetStatus: Gets the status of a channel
The implementation in apiAdapter properly wraps these operations with logging and metrics. The interface and its implementation are consistent with the adapter pattern being used.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Search for the api interface definition in the same package
ast-grep --pattern 'type api interface {
$$$
}'
# Search for any references to this interface
rg -A 5 'type api interface'
Length of output: 82
Script:
#!/bin/bash
# Let's try searching for the api type in the same directory with a case-insensitive search
rg -i "type.*api" router/batchrouter/asyncdestinationmanager/snowpipestreaming/
# Also search for any files that might define this API interface
fd -e go . router/batchrouter/asyncdestinationmanager/snowpipestreaming/ --exec cat {}
Length of output: 85907
Script:
#!/bin/bash
# Let's search for the api interface implementation in the same package
rg -A 10 'type api interface' router/batchrouter/asyncdestinationmanager/snowpipestreaming/
# Also search for the api adapter implementation
rg -A 10 'type apiAdapter struct' router/batchrouter/asyncdestinationmanager/snowpipestreaming/
Length of output: 191
warehouse/integrations/manager/manager.go (2)
92-92: LGTM! Consistent implementation
The implementation in NewWarehouseOperations matches the one in New, maintaining consistency in how both destination types are handled.
67-67: Verify snowflake manager's compatibility with SnowpipeStreaming
The code reuses the snowflake manager for both SNOWFLAKE and SnowpipeStreaming destination types. While this suggests shared functionality, we should verify that the snowflake manager can handle both modes correctly.
✅ Verification successful
Let me gather more information about how Snowflake manager handles different destination types.
Let me try a different approach to check the snowflake implementation.
Let me check one more aspect to verify how the snowflake manager handles different destination types.
Based on the gathered information, I can now provide the final response:
Snowflake manager safely handles both destination types
The snowflake manager implementation is destination-type agnostic and handles both SNOWFLAKE and SnowpipeStreaming types correctly. The manager's core functionality:
- Uses the same connection and configuration setup for both types
- Handles data loading and schema management uniformly
- Relies on common utilities for cloud provider and object storage configuration
- Maintains consistent warehouse operations regardless of the destination type
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify if snowflake manager has specific handling for SnowpipeStreaming
# Check for any specific handling of SnowpipeStreaming in snowflake package
echo "Checking snowflake package for SnowpipeStreaming handling..."
rg -l "SnowpipeStreaming" warehouse/integrations/snowflake/
# Check for any configuration or setup differences
echo "Checking for configuration differences..."
ast-grep --pattern 'if destType == warehouseutils.SnowpipeStreaming {
$$$
}'
Length of output: 364
Script:
#!/bin/bash
# Let's check for any differences in handling between SNOWFLAKE and SnowpipeStreaming
# Check snowflake manager implementation
echo "Checking snowflake manager implementation..."
rg -A 5 "type SnowflakeManager struct" warehouse/integrations/snowflake/
# Check initialization and configuration
echo "Checking initialization and configuration..."
rg -A 10 "func NewSnowflake\(" warehouse/integrations/snowflake/
# Check if there are any mode-specific configurations
echo "Checking for destination type specific handling..."
rg -A 3 "destType.*=.*SnowpipeStreaming" warehouse/integrations/snowflake/
rg -A 3 "DestinationType.*=.*SnowpipeStreaming" warehouse/integrations/snowflake/
# Check for any snowpipe specific methods or configurations
echo "Checking for snowpipe specific implementations..."
rg -l "snowpipe" warehouse/integrations/snowflake/
Length of output: 771
Script:
#!/bin/bash
# Let's examine the snowflake package more thoroughly
# First, list all files in the snowflake package to understand its structure
echo "Listing snowflake package files..."
fd . warehouse/integrations/snowflake/
# Check for any streaming-related implementations
echo "Checking for streaming implementations..."
rg -l "Stream" warehouse/integrations/snowflake/
# Look for configuration or setup related code
echo "Checking for configuration setup..."
rg -A 5 "Config" warehouse/integrations/snowflake/
# Check for any type-specific handling in the upload process
echo "Checking upload process..."
rg -A 5 "Upload" warehouse/integrations/snowflake/
# Check for any initialization differences
echo "Checking initialization..."
rg -A 5 "Init" warehouse/integrations/snowflake/
Length of output: 86754
Script:
#!/bin/bash
# Check for any differences in handling between SNOWFLAKE and SnowpipeStreaming
# Check the main snowflake implementation file
echo "Checking snowflake.go for destination type handling..."
rg -A 5 "type Snowflake struct" warehouse/integrations/snowflake/snowflake.go
# Check for any configuration differences based on destination type
echo "Checking for destination type specific configuration..."
rg -A 5 "DestinationType.*=.*" warehouse/integrations/snowflake/snowflake.go
# Check for any setup differences
echo "Checking Setup method for destination type handling..."
rg -A 10 "func.*Setup.*" warehouse/integrations/snowflake/snowflake.go
Length of output: 1214
router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming_test.go (2)
1-10: LGTM! Clean import organization.
The package declaration and imports are well-organized, following Go conventions with proper grouping of standard library, external, and internal packages.
109-110: Verify the implementation of tested functions.
The test file references two functions that should be implemented elsewhere:
findNewColumnsfunctiondestConfig.Decodemethod
Also applies to: 167-168
✅ Verification successful
Let me search for the findNewColumns function in the snowpipestreaming package specifically.
Both tested functions are properly implemented in the codebase
The verification confirms:
findNewColumnsfunction is implemented inrouter/batchrouter/asyncdestinationmanager/snowpipestreaming/channel.godestConfig.Decodemethod is implemented inrouter/batchrouter/asyncdestinationmanager/snowpipestreaming/types.go
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify the implementation of tested functions exists
# Search for findNewColumns function implementation
echo "Searching for findNewColumns implementation..."
ast-grep --pattern 'func findNewColumns($_, $_) {
$$$
}'
# Search for destConfig struct and Decode method
echo "Searching for destConfig struct and Decode method..."
ast-grep --pattern 'type destConfig struct {
$$$
}'
ast-grep --pattern 'func ($_) Decode($_) error {
$$$
}'
Length of output: 2336
Script:
#!/bin/bash
# Search for findNewColumns in snowpipestreaming package files
echo "Searching for findNewColumns in snowpipestreaming files..."
rg -l "snowpipestreaming" | xargs rg "findNewColumns" -A 5
# Also search for the function implementation using a more specific pattern
echo "Searching with specific pattern..."
ast-grep --pattern 'func findNewColumns(eventSchema []interface{}, snowPipeSchema []interface{}) []string {
$$$
}'
Length of output: 2472
router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api_integration_test.go (1)
32-47: LGTM! Clean and minimal mock implementations.
The mock types are well-designed for testing HTTP interactions.
warehouse/slave/worker.go (1)
348-348: Verify the impact of exporting HandleSchemaChange
The function has been made public (exported). Please ensure this change is intentional and verify if other packages need this functionality.
✅ Verification successful
Function export is justified and properly used
The function HandleSchemaChange is correctly exported as it's being used in another package (router/batchrouter/asyncdestinationmanager/snowpipestreaming/discards.go). The change from private to public scope is intentional and necessary for this cross-package usage.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Check if HandleSchemaChange is used in other packages
# Search for usage of HandleSchemaChange across the codebase
rg "HandleSchemaChange" --type go
# Search for potential test files that might need to be updated
rg "handleSchemaChange" --type go
Length of output: 744
router/batchrouter/handle_lifecycle.go (2)
214-216: LGTM! Configuration keys properly scoped to destination type.
The modification to include destination type in configuration keys for minIdleSleep, uploadFreq, and mainLoopFreq follows the established pattern and allows for destination-specific timing configurations.
Let's verify the configuration usage across destinations:
✅ Verification successful
Based on the search results, I can now generate the final response since I have sufficient information to verify the configuration changes:
Configuration changes properly implemented and used consistently
The destination-specific configuration keys for minIdleSleep, uploadFreq, and mainLoopFreq are correctly implemented and used throughout the codebase:
- The configuration values are properly loaded with destination type prefix in
handle_lifecycle.go - The values are used consistently in worker scheduling and execution logic
- Integration tests (especially
snowpipestreaming_test.go) demonstrate proper usage of destination-specific configurations - No hardcoded values found that should be using these configurations instead
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify configuration patterns and usage across destinations
# Expected: Find consistent usage of these configurations with destination type prefixes
# Search for minIdleSleep usage
rg "minIdleSleep" --type go -B 2 -A 2
# Search for uploadFreq configuration patterns
rg "uploadFreq" --type go -B 2 -A 2
# Search for mainLoopFreq usage
rg "mainLoopFreq" --type go -B 2 -A 2
# Check for any hardcoded values that should use these configurations
ast-grep --pattern 'time.Sleep($duration)'
Length of output: 34133
205-205: LGTM! Verify timeout configuration.
The addition of asyncUploadWorkerTimeout with a default of 10 seconds follows the established pattern. The configuration is properly prefixed with destination type for flexibility.
Let's verify if similar timeout configurations exist in other async destinations:
✅ Verification successful
The timeout configuration is appropriate and consistent with the codebase patterns
The asyncUploadWorkerTimeout with a default of 10 seconds aligns well with similar timeout patterns in the codebase:
- It follows the same configuration pattern as other BatchRouter timeouts
- The 10-second default is consistent with similar worker/upload timeouts:
asyncUploadTimeout(30 minutes for overall upload)pollStatusLoopSleep(10 seconds)minIdleSleep(2 seconds)uploadFreq(30 seconds)
The configuration is properly structured with destination-type specific override support (BatchRouter.<destType>.asyncUploadWorkerTimeout) and a global fallback (BatchRouter.asyncUploadWorkerTimeout).
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Check for similar worker timeout configurations across the codebase
# Expected: Find similar timeout patterns in other async destinations
# Search for worker timeout configurations
rg -A 2 "uploadWorkerTimeout|workerTimeout" --type go
# Search for default timeout values in async destinations
ast-grep --pattern 'config.GetReloadableDurationVar($_, $_, time.Second, $$$)'
Length of output: 136
Script:
#!/bin/bash
# Let's try a broader search for timeout configurations and worker-related settings
# Search for timeout configurations in batch router
rg -i "timeout.*=.*config" router/batchrouter/
# Search for worker-related configurations
rg -i "worker.*=.*config" router/batchrouter/
# Search for GetReloadableDurationVar usage
rg "GetReloadableDurationVar" -A 2 --type go
# Look for similar configuration patterns
rg "BatchRouter\." -A 2 --type go
Length of output: 42470
gateway/handle_lifecycle.go (1)
41-41: LGTM: Import addition is well-placed
The addition of the timeutil package import is correctly placed and aligns with the implementation changes.
processor/transformer/transformer.go (1)
540-542: LGTM! The implementation follows existing patterns.
The Snowpipe Streaming destination handling is properly integrated, following the same pattern as other warehouse destinations and reusing existing configuration values.
Let's verify the configuration documentation:
utils/misc/misc.go (1)
99-99: LGTM! Verify integration with destination handlers
The addition of "SNOWPIPE_STREAMING" to the batch destinations list is consistent with the PR's objective of adding Snowpipe Streaming support.
Let's verify that the new destination type is properly handled in the destination-specific code:
✅ Verification successful
Based on the comprehensive verification results, I can now generate the final response:
LGTM! Snowpipe Streaming integration is properly implemented
The addition of "SNOWPIPE_STREAMING" to the batch destinations list is well-integrated:
- Complete implementation found in
router/batchrouter/asyncdestinationmanager/snowpipestreaming/ - Proper destination type handling in warehouse integrations manager
- Comprehensive test coverage including integration tests
- Full implementation of required interfaces (Upload, Poll, GetUploadStats)
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify that SNOWPIPE_STREAMING is properly integrated with destination handlers
# Check for destination handler implementation
ast-grep --pattern 'case "SNOWPIPE_STREAMING": $$$'
# Check for any configuration structs or interfaces related to Snowpipe
rg -l "(?i)snowpipe.*(?:config|options|settings)"
# Check for any test files related to Snowpipe
fd -g "*test*.go" | xargs rg -l "(?i)snowpipe"
Length of output: 816
Script:
#!/bin/bash
# Let's verify the integration by checking the implementation details
# Check the main implementation file for proper handling
cat router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming.go
# Check if the destination type is properly handled in router configuration
rg -A 5 "type DestinationType string"
# Check for any switch/case statements handling destination types
ast-grep --pattern 'switch destType {
$$$
}'
Length of output: 26001
warehouse/utils/utils.go (2)
41-52: LGTM: Constants block is well-structured
The addition of SnowpipeStreaming constant follows the established naming convention and is appropriately placed within the warehouse destinations block.
520-521: LGTM: Consistent handling of SnowpipeStreaming case
The modification to include SnowpipeStreaming in the uppercase conversion logic is consistent with the existing pattern and correctly groups it with SNOWFLAKE.
warehouse/slave/worker_test.go (3)
Line range hint 964-975: LGTM! Well-structured and comprehensive test cases
The test implementation demonstrates good practices:
- Comprehensive coverage of data type conversions
- Well-organized table-driven tests
- Parallel test execution for better performance
- Clear test case names and expectations
Line range hint 1-975: LGTM! Excellent test structure and coverage
The test file demonstrates excellent testing practices:
- Logical organization using sub-tests
- Proper isolation of integration tests using docker containers
- Thorough error case coverage
- Comprehensive assertions
964-964: Verify the exported function's usage and documentation
The function has been made public (exported) by renaming from handleSchemaChange to HandleSchemaChange. This change requires verification of its usage across the codebase and proper documentation.
✅ Verification successful
Based on the results, I can now generate the final response:
Function export change is properly implemented and documented
The HandleSchemaChange function is:
- Properly documented in
warehouse/slave/worker.go - Used in both test files and production code
- Correctly exported as it's being used across packages (specifically in
router/batchrouter/asyncdestinationmanager/snowpipestreaming/discards.go)
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Check for function usage and documentation
# Test 1: Find all usages of HandleSchemaChange
rg -p "HandleSchemaChange"
# Test 2: Check if the function is documented in any package documentation
fd -e go -x grep -l "^// HandleSchemaChange" {}
Length of output: 725
gateway/handle.go (1)
491-491: LGTM! Verify time handling consistency.
The change to use gw.now() instead of time.Now() improves testability. This is a good practice as it allows for time mocking in tests.
Let's verify the consistency of timestamp handling across the codebase:
✅ Verification successful
Based on the search results, I can now generate the final response:
LGTM! Time handling is consistent across the codebase
The change to use gw.now() instead of time.Now() is consistent with the codebase's time handling patterns:
- The timestamp format
misc.RFC3339Milliis consistently used across the codebase for event timestamps - The gateway package already uses
gw.now()in other places for timestamp generation - The change maintains consistency with the existing pattern of dependency injection for time functions, which improves testability
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Check for direct time.Now() calls and timestamp format consistency
# Check for direct time.Now() calls in the gateway package
echo "Checking for direct time.Now() calls in gateway package..."
rg "time\.Now\(\)" --type go
# Check for RFC3339Milli format usage consistency
echo "Checking timestamp format consistency..."
rg "RFC3339|time\.RFC" --type go
Length of output: 66261
warehouse/integrations/snowflake/snowflake.go (1)
1403-1403: LGTM! Verify other usages of the exported function.
The change to use CalculateDataType with explicit Int64 access improves code reusability and type safety.
Let's verify other usages of this function across the codebase:
✅ Verification successful
Inconsistent usage of calculateDataType function across the codebase
The verification reveals that there are two different implementations:
calculateDataTypeinwarehouse/integrations/redshift/redshift.goCalculateDataTypeinwarehouse/integrations/snowflake/datatype_mapper.go
Let's verify if these implementations are similar and should be consolidated:
Let me try a different pattern to fetch the implementations:
The change to CalculateDataType is correctly implemented but reveals separate concerns
The verification shows that while both functions share a similar structure, they handle different warehouse-specific logic:
- Redshift's version handles string length constraints
- Snowflake's version handles numeric scale for float conversion
These implementations should remain separate as they address warehouse-specific type mapping requirements. The change to export Snowflake's function is appropriate as it's now being used by the snowpipe streaming module.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Find all usages of the calculateDataType/CalculateDataType function
# Search for both old and new function names to ensure complete migration
echo "Searching for old function name (calculateDataType)..."
rg "calculateDataType\("
echo "Searching for new function name (CalculateDataType)..."
rg "CalculateDataType\("
Length of output: 1186
Script:
#!/bin/bash
# Get implementation of both functions
echo "Redshift calculateDataType implementation:"
ast-grep --pattern 'func calculateDataType($_, $_) { $$$ }' warehouse/integrations/redshift/redshift.go
echo -e "\nSnowflake CalculateDataType implementation:"
ast-grep --pattern 'func CalculateDataType($_, $_) { $$$ }' warehouse/integrations/snowflake/datatype_mapper.go
Length of output: 417
Script:
#!/bin/bash
echo "Redshift calculateDataType implementation:"
rg -A 20 "^func calculateDataType" warehouse/integrations/redshift/redshift.go
echo -e "\nSnowflake CalculateDataType implementation:"
rg -A 20 "^func CalculateDataType" warehouse/integrations/snowflake/datatype_mapper.go
Length of output: 1277
warehouse/utils/reservedkeywords.go (1)
97-189: LGTM! The SNOWPIPE_STREAMING keywords are correctly defined.
The addition of SNOWPIPE_STREAMING keywords is appropriate as it matches the SNOWFLAKE keywords, which is expected since Snowpipe Streaming is a Snowflake feature using the same SQL dialect.
router/batchrouter/asyncdestinationmanager/snowpipestreaming/types.go (2)
32-43: Ensure Proper Initialization of Client Configuration Parameters
The client configuration parameters within the Manager struct's config field do not show default values in this snippet. Ensure that these fields are properly initialized elsewhere in the codebase. Uninitialized configuration settings can lead to runtime issues.
22-31: 🛠️ Refactor suggestion
Potential Confusion Between appConfig and config Fields in Manager Struct
The Manager struct contains both appConfig and config fields, which could lead to confusion due to their similar naming. Consider renaming the config field to something more specific, such as managerConfig, to improve clarity.
This is similar to a previous comment where conf was renamed to appConfig.
router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/model/model.go (1)
14-14: Consider defining types individually for better clarity.
Defining types individually rather than within a grouped type block can improve readability and make it easier to locate and maintain each type.
router/batchrouter/asyncdestinationmanager/snowpipestreaming/discards.go (1)
68-76: Function names should start with a verb for clarity
Consider renaming discardsTable and discardsSchema to getDiscardsTable and getDiscardsSchema to follow Go naming conventions, enhancing readability and consistency across the codebase.
router/batchrouter/asyncdestinationmanager/snowpipestreaming/channel.go (1)
229-248: Verify secure handling of authentication configuration
The code sets useKeyPairAuth to true, enforcing key pair authentication. Ensure that private keys and passphrases in destConf are securely managed and not exposed in logs or error messages throughout the codebase.
router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming.go (3)
463-466: Logic Clarification: Condition for Import Completion
The condition statusRes.Offset != info.Offset determines if the import is still in progress. Confirm that this logic correctly reflects the intended behavior, as it might be counterintuitive.
Please verify if the condition should be statusRes.Offset == info.Offset to indicate completion.
375-376: Typographical Error: Incorrect HTTP Status Code
The HTTP status code for a bad request should be http.StatusBadRequest, which is 400. However, in line 375, you might want to double-check if this status code aligns with the intended response in this context.
Ensure that returning http.StatusBadRequest is appropriate here, or if another status code like http.StatusInternalServerError would be more suitable.
236-265: Potential Performance Issue: Reading Large Files Without Limiting Buffer Size
When reading events from the file in eventsFromFile, the scanner buffer is set based on maxBufferCapacity. Ensure that maxBufferCapacity is appropriately sized to prevent potential memory issues with large files.
Run the following script to verify the usages and default value of maxBufferCapacity:
Ensure that maxBufferCapacity is set to a reasonable value and consider making it configurable if necessary.
✅ Verification successful
Buffer Size is Properly Configured and Limited
The verification shows that maxBufferCapacity is properly configured with a reasonable default value of 512KB and is implemented as a reloadable configuration variable. This is a safe and appropriate size for buffered scanning, as it:
- Has a sensible default (512KB)
- Is configurable at runtime through configuration
- Uses proper unit handling through
bytesize.KB - Is explicitly used to limit the scanner's buffer size
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify the default value and usages of maxBufferCapacity.
# Check for the default value setting
rg -A 2 'm\.config\.maxBufferCapacity' -- 'snowpipestreaming.go'
# Check for other usages of maxBufferCapacity
rg 'maxBufferCapacity' -- 'snowpipestreaming.go'
Length of output: 233
Script:
#!/bin/bash
# Let's first find the correct path to the file
fd snowpipestreaming.go
# After finding the file, let's check for maxBufferCapacity usage
fd snowpipestreaming.go --exec rg -A 2 'maxBufferCapacity'
# Let's also check for any configuration struct definitions
fd snowpipestreaming.go --exec ast-grep --pattern 'type $_ struct {
$$$
maxBufferCapacity $_
$$$
}'
Length of output: 508
router/batchrouter/handle_async.go (2)
18-18: Import statement added correctly
The addition of github.com/rudderlabs/rudder-go-kit/stats is appropriate and necessary for the updated functionality.
381-381: Standardize time retrieval using brt.now()
Replacing time.Now() with brt.now() enhances consistency in time handling across the codebase, allowing for easier testing and potential future adjustments to time retrieval mechanisms.
router/batchrouter/handle.go (2)
31-31: Verify the necessity of the new import asynccommon.
The import of asynccommon "github.com/rudderlabs/rudder-server/router/batchrouter/asyncdestinationmanager/common" has been added. Please ensure that this package is used in the code. Unused imports can clutter the codebase and introduce unnecessary dependencies.
44-44: Confirm usage of the timeutil package.
The timeutil package has been imported with github.com/rudderlabs/rudder-server/utils/timeutil. Verify that the utilities provided by this package are effectively utilized in the code. If it's only used for timeutil.Now, consider whether importing the entire package is necessary or if time.Now suffices.
router/batchrouter/asyncdestinationmanager/snowpipestreaming/apiadapter.go
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/apiadapter.go
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/model/model_test.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api_test.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming.go
Outdated
Show resolved
Hide resolved
router/batchrouter/asyncdestinationmanager/snowpipestreaming/snowpipestreaming.go
Show resolved
Hide resolved
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
3 times, most recently
from
9e6e054 to
2244a7d
Compare
router/batchrouter/asyncdestinationmanager/snowpipestreaming/internal/api/api.go
Show resolved
Hide resolved
| return getDiscardedRecordsFromEvent(tableEvent, channelResponse.SnowPipeSchema, info.tableName, formattedTS) | ||
| }) | ||
|
|
||
| insertReq := &model.InsertRequest{ |
There was a problem hiding this comment.
should we filter events that are there in discardInfos before creating insertReq since we know they are bound to fail?
There was a problem hiding this comment.
We don't need to filtering events. Discards are being generated at a property level. 1 event can have multiple discards, we set those properties as nil and send it to snowflake.
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
from
861991a to
8874345
Compare
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
2 times, most recently
from
01bd7f9 to
a245cc7
Compare
achettyiitr
force-pushed
the
feat.snowpipe-streaming
branch
from
a245cc7 to
db5364b
Compare
Description
Security