Merge branch 'master' into sync

* master: iter.go: error message typo correction (thanos-io#1376) Fix usage of $GOPATH in Makefile (thanos-io#1379) Moved Prometheus 2.11.1 and TSDB to 0.9.1 (thanos-io#1380) Store latest config hash and timestamp as metrics (thanos-io#1378) pkg/receive/handler.go: log errors (thanos-io#1372) receive: Hash-ring metrics (thanos-io#1363) receiver: avoid race of hashring (thanos-io#1371) feat compact: added readiness Prober (thanos-io#1297) Add changelog entry for S3 option (thanos-io#1361) Multipart features (thanos-io#1358) Added katacoda.yaml (thanos-io#1359) Remove deprecated option from example (thanos-io#1351) Move suggestion about admin API to appropriate place (thanos-io#1355)
paulfantom · Aug 7, 2019 · 3e43fbd · 3e43fbd
2 parents 9c6af21 + 343c04c
commit 3e43fbd
Show file tree

Hide file tree

Showing 60 changed files with 1,224 additions and 798 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -11,12 +11,24 @@ We use *breaking* word for marking changes that are not backward compatible (rel
 
 ## Unreleased.
 
+### Added
+
+- [#1358](https://github.com/thanos-io/thanos/pull/1358) Added `part_size` configuration option for HTTP multipart requests minimum part size for S3 storage type
+- [#1363](https://github.com/thanos-io/thanos/pull/1363) Thanos Receive now exposes `thanos_receive_hashring_nodes` and `thanos_receive_hashring_tenants` metrics to monitor status of hash-rings
+
+### Changed
+
+- [#1380](https://github.com/thanos-io/thanos/pull/1380) Upgraded important dependencies: Prometheus to 2.11.1 and TSDB to 0.9.1. Some changes affecting Querier:
+  - [ENHANCEMENT] Query performance improvement: Efficient iteration and search in HashForLabels and HashWithoutLabels. #5707
+  - [ENHANCEMENT] Optimize queries using regexp for set lookups. tsdb#602
+  - [BUGFIX] prometheus_tsdb_compactions_failed_total is now incremented on any compaction failure. tsdb#613
+  - [BUGFIX] PromQL: Correctly display {__name__="a"}. #5552
 - [#1338](https://github.com/thanos-io/thanos/pull/1338) Querier still warns on store API duplicate, but allows a single one from duplicated set. This is gracefully warn about the problematic logic and not disrupt immediately.
+- [#1297](https://github.com/improbable-eng/thanos/pull/1297) Added `/-/ready` and `/-/healthy` endpoints to Thanos compact.
 
 ### Fixed
 
 - [#1327](https://github.com/thanos-io/thanos/pull/1327) `/series` API end-point now properly returns an empty array just like Prometheus if there are no results
-
 - [#1302](https://github.com/thanos-io/thanos/pull/1302) Thanos now efficiently reuses HTTP keep-alive connections
 
 ## [v0.6.0](https://github.com/thanos-io/thanos/releases/tag/v0.6.0) - 2019.07.18
@@ -59,7 +71,7 @@ The other `type` you can use is `JAEGER` now. The `config` keys and values are J
 
 ### Changed
 
-- [#1284](https://github.com/thanos-io/thanos/pull/1284) Add support for multiple label-sets in Info gRPC service. 
+- [#1284](https://github.com/thanos-io/thanos/pull/1284) Add support for multiple label-sets in Info gRPC service.
 This deprecates the single `Labels` slice of the `InfoResponse`, in a future release backward compatible handling for the single set of Labels will be removed. Upgrading to v0.6.0 or higher is advised.
 *breaking* If you run have duplicate queries in your Querier configuration with hierarchical federation of multiple Queries this PR makes Thanos Querier to detect this case and block all duplicates. Refer to 0.6.1 which at least allows for single replica to work.
 

diff --git a/Makefile b/Makefile
@@ -5,7 +5,7 @@ DOCKER_IMAGE_NAME ?= thanos
 DOCKER_IMAGE_TAG  ?= $(subst /,-,$(shell git rev-parse --abbrev-ref HEAD))-$(shell date +%Y-%m-%d)-$(shell git rev-parse --short HEAD)
 
 TMP_GOPATH        ?= /tmp/thanos-go
-GOBIN             ?= ${GOPATH}/bin
+GOBIN             ?= $(firstword $(subst :, ,$GOPATH))/bin
 GO111MODULE       ?= on
 export GO111MODULE
 GOPROXY           ?= https://proxy.golang.org

diff --git a/cmd/thanos/compact.go b/cmd/thanos/compact.go
@@ -15,18 +15,20 @@ import (
 	"github.com/go-kit/kit/log"
 	"github.com/go-kit/kit/log/level"
 	"github.com/oklog/run"
-	opentracing "github.com/opentracing/opentracing-go"
+	"github.com/opentracing/opentracing-go"
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/tsdb"
 	"github.com/thanos-io/thanos/pkg/block"
 	"github.com/thanos-io/thanos/pkg/block/metadata"
 	"github.com/thanos-io/thanos/pkg/compact"
 	"github.com/thanos-io/thanos/pkg/compact/downsample"
+	"github.com/thanos-io/thanos/pkg/component"
 	"github.com/thanos-io/thanos/pkg/objstore"
 	"github.com/thanos-io/thanos/pkg/objstore/client"
+	"github.com/thanos-io/thanos/pkg/prober"
 	"github.com/thanos-io/thanos/pkg/runutil"
-	kingpin "gopkg.in/alecthomas/kingpin.v2"
+	"gopkg.in/alecthomas/kingpin.v2"
 )
 
 var (
@@ -49,7 +51,7 @@ func (cs compactionSet) String() string {
 	return strings.Join(result, ", ")
 }
 
-// levels returns set of compaction levels not higher than specified max compaction level
+// levels returns set of compaction levels not higher than specified max compaction level.
 func (cs compactionSet) levels(maxLevel int) ([]int64, error) {
 	if maxLevel >= len(cs) {
 		return nil, errors.Errorf("level is bigger then default set of %d", len(cs))
@@ -62,13 +64,14 @@ func (cs compactionSet) levels(maxLevel int) ([]int64, error) {
 	return levels, nil
 }
 
-// maxLevel returns max available compaction level
+// maxLevel returns max available compaction level.
 func (cs compactionSet) maxLevel() int {
 	return len(cs) - 1
 }
 
-func registerCompact(m map[string]setupFunc, app *kingpin.Application, name string) {
-	cmd := app.Command(name, "continuously compacts blocks in an object store bucket")
+func registerCompact(m map[string]setupFunc, app *kingpin.Application) {
+	comp := component.Compact
+	cmd := app.Command(comp.String(), "continuously compacts blocks in an object store bucket")
 
 	haltOnError := cmd.Flag("debug.halt-on-error", "Halt the process if a critical compaction error is detected.").
 		Hidden().Default("true").Bool()
@@ -110,7 +113,7 @@ func registerCompact(m map[string]setupFunc, app *kingpin.Application, name stri
 	compactionConcurrency := cmd.Flag("compact.concurrency", "Number of goroutines to use when compacting groups.").
 		Default("1").Int()
 
-	m[name] = func(g *run.Group, logger log.Logger, reg *prometheus.Registry, tracer opentracing.Tracer, _ bool) error {
+	m[comp.String()] = func(g *run.Group, logger log.Logger, reg *prometheus.Registry, tracer opentracing.Tracer, _ bool) error {
 		return runCompact(g, logger, reg,
 			*httpAddr,
 			*dataDir,
@@ -125,7 +128,7 @@ func registerCompact(m map[string]setupFunc, app *kingpin.Application, name stri
 				compact.ResolutionLevel5m:  time.Duration(*retention5m),
 				compact.ResolutionLevel1h:  time.Duration(*retention1h),
 			},
-			name,
+			comp,
 			*disableDownsampling,
 			*maxCompactionLevel,
 			*blockSyncConcurrency,
@@ -147,7 +150,7 @@ func runCompact(
 	wait bool,
 	generateMissingIndexCacheFiles bool,
 	retentionByResolution map[compact.ResolutionLevel]time.Duration,
-	component string,
+	component component.Component,
 	disableDownsampling bool,
 	maxCompactionLevel int,
 	blockSyncConcurrency int,
@@ -168,12 +171,18 @@ func runCompact(
 
 	downsampleMetrics := newDownsampleMetrics(reg)
 
+	readinessProber := prober.NewProber(component, logger, prometheus.WrapRegistererWithPrefix("thanos_", reg))
+	// Initiate default HTTP listener providing metrics endpoint and readiness/liveness probes.
+	if err := defaultHTTPListener(g, logger, reg, httpBindAddr, readinessProber); err != nil {
+		return errors.Wrap(err, "create readiness prober")
+	}
+
 	confContentYaml, err := objStoreConfig.Content()
 	if err != nil {
 		return err
 	}
 
-	bkt, err := client.NewBucket(logger, confContentYaml, reg, component)
+	bkt, err := client.NewBucket(logger, confContentYaml, reg, component.String())
 	if err != nil {
 		return err
 	}
@@ -318,11 +327,8 @@ func runCompact(
 		cancel()
 	})
 
-	if err := metricHTTPListenGroup(g, logger, reg, httpBindAddr); err != nil {
-		return err
-	}
-
 	level.Info(logger).Log("msg", "starting compact node")
+	readinessProber.SetReady()
 	return nil
 }
 

diff --git a/cmd/thanos/main.go b/cmd/thanos/main.go
@@ -31,6 +31,7 @@ import (
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
 	"github.com/prometheus/common/version"
+	"github.com/thanos-io/thanos/pkg/prober"
 	"github.com/thanos-io/thanos/pkg/runutil"
 	"github.com/thanos-io/thanos/pkg/tracing"
 	"github.com/thanos-io/thanos/pkg/tracing/client"
@@ -73,7 +74,7 @@ func main() {
 	registerStore(cmds, app, "store")
 	registerQuery(cmds, app, "query")
 	registerRule(cmds, app, "rule")
-	registerCompact(cmds, app, "compact")
+	registerCompact(cmds, app)
 	registerBucket(cmds, app, "bucket")
 	registerDownsample(cmds, app, "downsample")
 	registerReceive(cmds, app, "receive")
@@ -122,7 +123,7 @@ func main() {
 	)
 
 	prometheus.DefaultRegisterer = metrics
-	// Memberlist uses go-metrics
+	// Memberlist uses go-metrics.
 	sink, err := gprom.NewPrometheusSink()
 	if err != nil {
 		fmt.Fprintln(os.Stderr, errors.Wrapf(err, "%s command failed", cmd))
@@ -311,6 +312,7 @@ func defaultGRPCServerOpts(logger log.Logger, reg *prometheus.Registry, tracer o
 	return append(opts, grpc.Creds(credentials.NewTLS(tlsCfg))), nil
 }
 
+// TODO Remove once all components are migrated to the new defaultHTTPListener.
 // metricHTTPListenGroup is a run.Group that servers HTTP endpoint with only Prometheus metrics.
 func metricHTTPListenGroup(g *run.Group, logger log.Logger, reg *prometheus.Registry, httpBindAddr string) error {
 	mux := http.NewServeMux()
@@ -330,3 +332,27 @@ func metricHTTPListenGroup(g *run.Group, logger log.Logger, reg *prometheus.Regi
 	})
 	return nil
 }
+
+// defaultHTTPListener starts a run.Group that servers HTTP endpoint with default endpoints providing Prometheus metrics,
+// profiling and liveness/readiness probes.
+func defaultHTTPListener(g *run.Group, logger log.Logger, reg *prometheus.Registry, httpBindAddr string, readinessProber *prober.Prober) error {
+	mux := http.NewServeMux()
+	registerMetrics(mux, reg)
+	registerProfile(mux)
+	readinessProber.RegisterInMux(mux)
+
+	l, err := net.Listen("tcp", httpBindAddr)
+	if err != nil {
+		return errors.Wrap(err, "listen metrics address")
+	}
+
+	g.Add(func() error {
+		level.Info(logger).Log("msg", "listening for metrics", "address", httpBindAddr)
+		readinessProber.SetHealthy()
+		return errors.Wrap(http.Serve(l, mux), "serve metrics")
+	}, func(err error) {
+		readinessProber.SetNotHealthy(err)
+		runutil.CloseWithLogOnErr(logger, l, "metric listener")
+	})
+	return nil
+}
diff --git a/docs/components/store.md b/docs/components/store.md
@@ -12,7 +12,6 @@ It keeps a small amount of information about all remote blocks on local disk and
 ```bash
 $ thanos store \
     --data-dir        "/local/state/data/dir" \
-    --cluster.peers    "thanos-cluster.example.org" \
     --objstore.config-file "bucket.yml"
 ```
 

diff --git a/docs/storage.md b/docs/storage.md
@@ -62,6 +62,7 @@ config:
     insecure_skip_verify: false
   trace:
     enable: false
+  part_size: 0
 ```
 
 At a minimum, you will need to provide a value for the `bucket`, `endpoint`, `access_key`, and `secret_key` keys. The rest of the keys are optional.
@@ -74,6 +75,8 @@ You can configure the timeout settings for the HTTP client by setting the `http_
 
 Please refer to the documentation of [the Transport type](https://golang.org/pkg/net/http/#Transport) in the `net/http` package for detailed information on what each option does.
 
+`part_size` is specified in bytes and refers to the minimum file size used for multipart uploads, as some custom S3 implementations may have different requirements. A value of `0` means to use a default 128 MiB size.
+
 For debug and testing purposes you can set
 
 * `insecure: true` to switch to plain insecure HTTP instead of HTTPS

diff --git a/go.mod b/go.mod
@@ -4,6 +4,7 @@ require (
 	cloud.google.com/go v0.34.0
 	github.com/Azure/azure-storage-blob-go v0.7.0
 	github.com/NYTimes/gziphandler v1.1.1
+	github.com/alecthomas/template v0.0.0-20190718012654-fb15b899a751 // indirect
 	github.com/armon/go-metrics v0.0.0-20180917152333-f0300d1749da
 	github.com/cespare/xxhash v1.1.0
 	github.com/fatih/structtag v1.0.0
@@ -33,8 +34,8 @@ require (
 	github.com/pkg/errors v0.8.1
 	github.com/prometheus/client_golang v1.0.0
 	github.com/prometheus/common v0.6.0
-	github.com/prometheus/prometheus v2.9.2+incompatible
-	github.com/prometheus/tsdb v0.8.0
+	github.com/prometheus/prometheus v0.0.0-20190710134608-e5b22494857d
+	github.com/prometheus/tsdb v0.9.1
 	github.com/uber-go/atomic v1.4.0 // indirect
 	github.com/uber/jaeger-client-go v2.16.0+incompatible
 	github.com/uber/jaeger-lib v2.0.0+incompatible