Re-enqueue 429 requests if there are multiple query-schedulers #5496
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
yeya24
reviewed
Aug 3, 2023
pkg/frontend/v2/frontend.go
Outdated
| @@ -192,6 +194,8 @@ func (f *Frontend) RoundTripGRPC(ctx context.Context, req *httpgrpc.HTTPRequest) | |||
| // even if this goroutine goes away due to client context cancellation. | |||
| enqueue: make(chan enqueueResult, 1), | |||
| response: make(chan *frontendv2pb.QueryResultRequest, 1), | |||
|
|
|||
| retryOnTooManyOutstandingRequests: f.schedulerWorkers.getWorkersCount() > 0, | |||
There was a problem hiding this comment.
If multiple query schedulers have queues full, then retry may make queueing even worse? A request will retry forever until it is not discarded by the queue.
There was a problem hiding this comment.
I think the retry can still be useful. Let's add a feature flag?
There was a problem hiding this comment.
retries := f.cfg.WorkerConcurrency + 1 limits the maximum number of retries.
A feature flag sounds good.
|
@damnever Can you fix DCO? |
Signed-off-by: Xiaochao Dong (@damnever) <the.xcdong@gmail.com>
Signed-off-by: Xiaochao Dong (@damnever) <the.xcdong@gmail.com>
yeya24
approved these changes
Aug 4, 2023
wenxu1024
pushed a commit
to wenxu1024/cortex
that referenced
this pull request
Aug 9, 2023
…xproject#5496) * Re-enqueue 429 requests if there are multiple query-schedulers Signed-off-by: Xiaochao Dong (@damnever) <the.xcdong@gmail.com> * Add feature flag Signed-off-by: Xiaochao Dong (@damnever) <the.xcdong@gmail.com> --------- Signed-off-by: Xiaochao Dong (@damnever) <the.xcdong@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What this PR does:
The current implementation cannot ensure that one tenant's requests are evenly distributed among all schedulers. As a result, some scheduler queues may be full, while others still have some room left.
This PR attempts to retry 429 requests randomly if there are multiple query-schedulers. We may need a more sophisticated approach to effectively address this issue.
Checklist
CHANGELOG.mdupdated - the order of entries should be[CHANGE],[FEATURE],[ENHANCEMENT],[BUGFIX]