-
Notifications
You must be signed in to change notification settings - Fork 1.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[extension/ballastextension] Deprecate memory_ballast extension #8803
[extension/ballastextension] Deprecate memory_ballast extension #8803
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #8803 +/- ##
=======================================
Coverage 91.47% 91.47%
=======================================
Files 320 320
Lines 17187 17187
=======================================
Hits 15722 15722
Misses 1165 1165
Partials 300 300 ☔ View full report in Codecov by Sentry. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @mx-psi for pushing this and getting us closer to GOMEMLIMIT !
Not sure if this was already discussed, the ballast extension has been used in conjunction with the memory-limiter processor to control GC;
With GOMEMLIMIT
do we need to change the behavior of memory-limiter ?
The memory-limiter currently provides 2 main tasks:
- signal the pipeline to drop or stop accepting new data
- call GC explicitly
With GOMEMLIMIT
we should no longer worry about explicitly calling GC anymore, however, we might still need to keep the drop or stop accepting data?
I also understand that @dmitryax suggested to move the memory-limiter as a receiver helper.
@mx-psi would be good to look into the memory limiter as well. |
@bogdandrutu @dloucasfx There are a couple dependencies on the memory ballast in other parts of the Collector:
IMO, we should first
After these three happen, we can think about changing the behavior of other parts of the Collector that depend on the memory ballast, including the memory limiter (if the receiver part) and the metrics. Does that make sense? Do you think there is specific parts we should do now as part of this PR? |
This PR was marked stale due to lack of activity. It will be closed in 14 days. |
@bogdandrutu @dloucasfx can you take another look? |
This PR was marked stale due to lack of activity. It will be closed in 14 days. |
Looks like we are first going to enable |
Description:
Based on user reports on open-telemetry/opentelemetry-helm-charts/issues/891 and the discussion on #8343, we can deprecate the memory ballast extension in favor of using
GOMEMLIMIT
. This PR:GOMEMLIMIT
with the same approach as in the Helm chart (80% of memory limit)Once this PR is accepted, open-telemetry/opentelemetry-helm-charts/issues/891 can move ahead with enabling
useGOMEMLIMIT
by default on the Helm chart.Other issues will be opened for opentelemetry.io, the Opentelemetry Operator and other parts of the OpenTelemetry project to remove references to this extension once the PR is merged.
No explicit timeline is given for removal of the extension.
Link to tracking Issue: Updates #8343