fix(sdk-trace-base): Export processed spans while exporter failed #4287
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
Zirak
commented
Nov 10, 2023
| @@ -221,7 +221,7 @@ export abstract class BatchSpanProcessorBase<T extends BufferConfig> | |||
| const flush = () => { | |||
| this._isExporting = true; | |||
| this._flushOneBatch() | |||
| .then(() => { | |||
| .finally(() => { | |||
| this._isExporting = false; | |||
There was a problem hiding this comment.
Both the finally and catch set _isExporting. Initially I removed the assignment from the catch, but that gave me slight nervous jitters
While the exporter deals with a batch of spans, new spans may come in and wait to be exported. As previously implemented, a successful export would notice these waiting spans, triggering a renewed timer check, but not so for an unsuccessful export. The result was that, prior to this commit, a failing export may end up in a situation where no further spans will be exported. This is due to the behaviour of `_addToBuffer` when the queue is full: Imagine an export which fails after a long timeout (because of, for instance, network troubles). While the connection waits to be timed out, the span queue fills up. Once completely full, no new calls to recheck the timer will be done. On its own, this behaviour is fine. When combined with the patched bug, this leads to a rather confusing case where the exporter never tries exporting.
Zirak
force-pushed
the
export-queue-filled
branch
from
e6d8d3e
to
2f1d7a5
Compare
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## main #4287 +/- ##
==========================================
+ Coverage 92.20% 92.21% +0.01%
==========================================
Files 336 336
Lines 9515 9515
Branches 2017 2017
==========================================
+ Hits 8773 8774 +1
+ Misses 742 741 -1
|
|
This PR is stale because it has been open 60 days with no activity. Remove stale label or comment or this will be closed in 14 days. |
pichlermarc
approved these changes
Jan 24, 2024
pichlermarc
added
bug
Zirak
added a commit
to Zirak/opentelemetry-js
that referenced
this pull request
Sep 14, 2024
…en-telemetry#4287) * fix(sdk-trace-base): Export processed spans while exporter failed While the exporter deals with a batch of spans, new spans may come in and wait to be exported. As previously implemented, a successful export would notice these waiting spans, triggering a renewed timer check, but not so for an unsuccessful export. The result was that, prior to this commit, a failing export may end up in a situation where no further spans will be exported. This is due to the behaviour of `_addToBuffer` when the queue is full: Imagine an export which fails after a long timeout (because of, for instance, network troubles). While the connection waits to be timed out, the span queue fills up. Once completely full, no new calls to recheck the timer will be done. On its own, this behaviour is fine. When combined with the patched bug, this leads to a rather confusing case where the exporter never tries exporting. * fix(changelog): add entry --------- Co-authored-by: Marc Pichler <marc.pichler@dynatrace.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.
Which problem is this PR solving?
While the exporter deals with a batch of spans, new spans may come in and wait to be exported. As previously implemented, a successful export would notice these waiting spans, triggering a renewed timer check, but not so for an unsuccessful export.
The result was that, prior to this commit, a failing export may end up in a situation where no further spans will be exported. This is due to the behaviour of
_addToBufferwhen the queue is full: Imagine an export which fails after a long timeout (because of, for instance, network troubles). While the connection waits to be timed out, the span queue fills up. Once completely full, no new calls to recheck the timer will be done. On its own, this behaviour is fine. When combined with the patched bug, this leads to a rather confusing case where the exporter never tries exporting.Short description of the changes
This PR makes sure these spans are acknowledged and handled, even when exporting fails.
@pichlermarc you've mentioned working on refactoring the exporter; here's an interesting edge case.
Type of change
Please delete options that are not relevant.
How Has This Been Tested?
Checklist: