[receiver/k8scluster] Do not keep metrics in memory #24769
Merged
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Outdated suggestions cannot be applied.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.
Construct metrics on every scrape instead of keeping them in memory and copy them with modifications on every scrape. We keep k8s objects in cache anyway, so we can use that instead of the pre-built metrics. This reduces RAM utilization.
This also allows us to extract a metrics builder instance and not create it on every scrape. This is the recommended approach that all other receivers follow. It ensures that any warnings defined in the metadata.yaml will be displayed only once, not on every scrape interval.
Memory utilization before/after the change on a cluster with 100 nodes and 1k pods:
