Flush usages data after publishOperation stream is closed #9444
Conversation
This file contains hidden or 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
maggie-lou
force-pushed
the
flush_usage_later
branch
from
699d43b to
cf6a6f5
Compare
vanja-p
approved these changes
May 23, 2025
bduffany
reviewed
May 27, 2025
Comment on lines
304
to
309
| // The scheduler attempts to prevent concurrent execution updates via | ||
| // task leasing, but this is not currently 100% reliable. If execution | ||
| // progress appears to restart, ignore any future updates. | ||
| if lastStage == int64(repb.ExecutionStage_COMPLETED) && event.GetStage() != int64(repb.ExecutionStage_COMPLETED) { | ||
| break | ||
| } |
There was a problem hiding this comment.
FYI we might (rarely) see some sequences like
(attempt 1) COMPLETED (execution completed)
(attempt 2) EXECUTING
(attempt 1) COMPLETED (recycling completed)
So we would drop the second COMPLETED event in this case. But it probably happens rarely enough that it's not worth worrying about for now.
bduffany
reviewed
May 27, 2025
| @@ -333,17 +337,70 @@ func (s *ExecutionServer) updateExecution(ctx context.Context, executionID strin | |||
| // Backwards-compatible fill of the execution with the ExecutionSummary for | |||
| // now. The ExecutionSummary will be removed completely in the future. | |||
| if statsUnset(md) { | |||
| if decodedMetadata, err := decodeMetadataFromExecutionSummary(executeResponse); err == nil { | |||
| if decodedMetadata, _ := decodeMetadataFromExecutionSummary(executeResponse); decodedMetadata != nil { | |||
There was a problem hiding this comment.
IIRC this was done in a different PR?
| @@ -440,14 +443,11 @@ type publishTest struct { | |||
| func testExecuteAndPublishOperation(t *testing.T, test publishTest) { | |||
There was a problem hiding this comment.
It would be good to add a test that does 2 different PublishOperation streams for the same execution, simulating a retry of a broken stream.
maggie-lou
force-pushed
the
flush_usage_later
branch
from
cf6a6f5 to
6b3985e
Compare
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.
Rather than flushing the data when a COMPLETED operation is published, publish when the client requests to close the publishOperation stream. This guarantees that all executions data has been collected (including cleanup data, which may come after an initial COMPLETED operation) before the data is flushed.