ratelimitmap: init maps before metrics collection #39361
Merged
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
Currently, the ratelimit map is initialized by the daemon init logic that is calling `ratelimitmap.InitMaps`. This commit refactors the ratelimitmap cell to init the map with a Hive lifecycle hook. Note: this is a preparation to do the same with the ratelimit metrics map where this refactoring is required to prevent races. Signed-off-by: Marco Hofstetter <marco.hofstetter@isovalent.com>
Currently, starting the Cilium Agent can result in the following warning when collecting metrics on the ratelimit metrics map starts before the daemon initialized the ratelimit metrics map. ``` 2025-05-06T10:17:39.944482772Z time="2025-05-06T10:17:39.941569631Z" level=warning msg="Failed to read ratelimit metrics from BPF map" source="/go/src/github.com/cilium/cilium/pkg/maps/ratelimitmap/ratelimitmap.go:203" error="loading pinned map /sys/fs/bpf/tc/globals/cilium_ratelimit_metrics: no such file or directory" subsys=ratelimit-map ``` The reason is that the map gets initialized via the daemon init logic (that calls `ratelimitmap.InitMaps` - while the metrics collector gets registered via Hive lifecycle. Therefore, this commit initializes the ratelimit metrics map with a Hive lifecycle hook and lets the metrics registration depend on the map. This way it's ensured that the map is created when the metrics collection starts. Signed-off-by: Marco Hofstetter <marco.hofstetter@isovalent.com>
cb1bc60 to
c591efb
Compare
|
/test |
joamaki
approved these changes
May 7, 2025
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Outdated suggestions cannot be applied.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.
Currently, starting the Cilium Agent can result in the following
warning when collecting metrics on the ratelimit metrics map starts
before the daemon initialized the ratelimit metrics map.
The reason is that the map gets initialized via the daemon init logic
(that calls
ratelimitmap.InitMaps- while the metrics collectorgets registered via Hive lifecycle.
Therefore, this commit initializes the ratelimit metrics map
with a Hive lifecycle hook and lets the metrics registration depend
on the map. This way it's ensured that the map is created when
the metrics collection starts.
Seen on #39357 in the ci-ginkgo tests (e.g. https://github.com/cilium/cilium/actions/runs/14856075257/job/41713009262). It's likely that the change of Daemon dependencies in that PR lead to exposing this issue. But the actual issue exists in the current codebase.
Agent logs from a sysdump
-> daemon init starts after ratelimit metrics collection