Bugfix: deduplicate concurrent resolver cache requests with singleflight.

[twoGiants]

Changes

Fix race condition in resolver cache where concurrent requests for the same resource bypass the cache and independently resolve upstream (cache stampede / thundering herd).

What changed

Before this change, cache lookup (Get) and storage (Add) were separate operations exposed on resolverCache, and the orchestration lived in GetFromCacheOrResolve in operations.go.

Now the entire check-or-resolve flow is a single GetCachedOrResolveFromRemote method on resolverCache that wraps the resolve callback with golang.org/x/sync/singleflight. Only one in-flight resolution proceeds per cache key; all concurrent callers share its result. The separate Get, Add, Remove, and GetFromCacheOrResolve public methods are removed — the cache surface is now just GetCachedOrResolveFromRemote and Clear.

Other changes

• Use strings.Builder in generateCacheKey instead of repeated string concatenation
• Remove unused resolverType parameter from ShouldUse
• Rework e2e tests to verify caching by counting actual registry GET requests (via container logs and registry /metrics endpoint) instead of checking resolver log messages
• Add multi-replica resolver e2e test (4 replicas, 80 TaskRuns) to validate singleflight works per-replica
• Allow -run flag to bypass category filtering in TestMain so individual tests can run in isolation
• Minor markdown lint fixes in test/README.md

@vdemeester also found and fixed a small bug in the e2e test annotations. I wasn't able to run tests in isolation. Now the little fix does the magic 🐱.

Fixes #9364

Submitter Checklist

As the author of this PR, please check off the items in this checklist:

• Has Docs if any changes are user facing, including updates to minimum requirements e.g. Kubernetes version bumps
• Has Tests included if any functionality added or changed
• pre-commit Passed
• Follows the commit message standard
• Meets the Tekton contributor standards (including functionality, content, code)
• Has a kind label. You can add one by adding a comment on this PR that contains /kind <type>. Valid types are bug, cleanup, design, documentation, feature, flake, misc, question, tep
• Release notes block below has been updated with any user facing changes (API changes, bug fixes, changes requiring upgrade notices or deprecation warnings). See some examples of good release notes.
• Release notes contains the string "action required" if the change requires additional action from users switching to the new release

Release Notes

Fix resolver cache race condition causing duplicate upstream pulls under concurrent load.

[tekton-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
To complete the pull request process, please ask for approval from twogiants after the PR has been reviewed.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[twoGiants]

@waveywaves check out the fix! 😺

[aThorp96]

Haven't gone through the tests yet but the fix looks good to me!

[twoGiants]

/retest

[vdemeester]

Two comments

[waveywaves]

/retest

[twoGiants]

We're good, take another look: @vdemeester @aThorp96

I updated the code, answered and marked the comments as resolved.