mirror of
https://github.com/cozystack/cozystack.git
synced 2026-10-30 08:37:31 +09:00
## What this PR does The lineage mutating webhook walks the owner chain (Pod, StatefulSet, HelmRelease, application CR) on every admission, with one uncached dynamic GET per level, and sibling Pods of the same parent repeat the same lookups. On a multi-tenant cluster this put `lineage.cozystack.io` at 300-500ms p99 in `apiserver_admission_webhook_admission_duration_seconds`, the largest share of admission latency. Owners are now cached in-process for 60 seconds, and only successful lookups are stored. A failed lookup (RESTMapping, NotFound, timeout) leaves the walk without an owner, so the object is admitted with no lineage labels and stays unlabeled until its next UPDATE runs the walk again. Secrets, Services and PVCs can go a long time between updates, so a cached failure would leave every object admitted within the TTL unlabeled for that long. The aggregated cozystack-api makes such failures routine, since discovery of `apps.cozystack.io` fails while it rolls. The one staleness left is the scheduling class, which is read from the application's spec: a change to it reaches newly admitted Pods up to 60 seconds later. The cache is not swept below 4096 entries, so up to that many owner objects stay in memory for the life of the process. Above that, each insert drops the expired entries among a random sample of 64, which keeps the write lock short. Non-Pod objects that already carry every lineage label, `internal.cozystack.io/tenantresource` included, skip the walk, because the objectSelector still matches an UPDATE whose old object lacks the label. Pods go through regardless, as each Pod admission also sets the scheduling class. ### Screenshots Not a UI change. ### Downstream repositories The diff is Go under `pkg/lineage` and `internal/lineagecontrollerwebhook` only: no label, annotation, CRD, values key, `hack/` file or telemetry metric is added or renamed, so no entry of the trigger map applies. - [x] No downstream repository is affected by this change - [ ] [cozystack/website](https://github.com/cozystack/website) - follow-up: - [ ] [cozystack/terraform-provider-cozystack](https://github.com/cozystack/terraform-provider-cozystack) - follow-up: - [ ] [cozystack/ansible-cozystack](https://github.com/cozystack/ansible-cozystack) - follow-up: - [ ] [cozystack/ccp](https://github.com/cozystack/ccp) - follow-up: - [ ] [cozystack/talm](https://github.com/cozystack/talm) - follow-up: - [ ] [cozystack/cozyhr](https://github.com/cozystack/cozyhr) - follow-up: - [ ] [cozystack/cozy-proxy](https://github.com/cozystack/cozy-proxy) - follow-up: - [ ] [cozystack/cozystack-telemetry-server](https://github.com/cozystack/cozystack-telemetry-server) - follow-up: - [ ] [cozystack/external-apps-example](https://github.com/cozystack/external-apps-example) - follow-up: - [ ] [cozystack/examples](https://github.com/cozystack/examples) - follow-up: - [ ] [cozystack/community](https://github.com/cozystack/community) - follow-up: ### Release note ```release-note perf(lineage): the lineage mutating webhook caches owner lookups for 60 seconds and skips the owner walk for non-Pod objects that already carry lineage labels, cutting admission latency on clusters with many tenants. ``` <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **New Features** * Lineage and owner lookups now use a time-limited in-memory cache, reducing repeated API requests during admission processing. Scheduling-class updates may take up to 60 seconds to appear on new Pods. * Non-Pod objects with complete lineage labels are allowed without additional owner lookup or mutation. * **Tests** * Added coverage for cache hits, misses, expiration, and retrying lookups after errors. <!-- end of auto-generated comment: release notes by coderabbit.ai -->