fix(prod): raise pipelines cpu request 500m→900m

[sumit-badsara]

cpu p95 865m = 173% of the old 500m request. Completes the pipelines right-size (memory was #32). Limit kept at 2 cores for bursts.

🤖 Generated with Claude Code

---

Note

Low Risk
Single production Helm values tweak for resource requests on an in-cluster private service; no app logic or exposure changes.

Overview
Raises the prod pipelines Deployment CPU request from 500m to 900m in k8s/prod/values.yaml, matching observed p95 ~865m (~173% of the old request). This follows the earlier memory right-size and aligns scheduling/HPA utilization with actual load.

CPU limit stays at 2 for burst headroom; memory requests/limits are unchanged.

^{Reviewed by Cursor Bugbot for commit 1077f79. Bugbot is set up for automated code reviews on this repo. Configure here.}

Greptile Summary

This PR raises the production CPU request for the pipelines workload.

• Updates k8s/prod/values.yaml from 500m to 900m.
• Keeps the CPU limit at 2 cores.
• Leaves the memory request and limit unchanged.

Confidence Score: 5/5

The change looks mergeable after a small cleanup to the resource comment.

• The Kubernetes value is a valid CPU request.
• The chart passes the resources block through directly.
• The only issue found is stale text next to the updated value.

k8s/prod/values.yaml

Important Files Changed

Filename	Overview
k8s/prod/values.yaml	Raises the production CPU request to 900m; the nearby comment still describes the old 500m default.

_{Reviews (1): Last reviewed commit: "fix(prod): raise pipelines cpu request 5..." | Re-trigger Greptile}

Greptile also left 1 inline comment on this PR.

Context used:

• Rule used - Ensure that the confidence score is always within ... (source)

Learned From
websentry-ai/ai-gateway-data#448

[greptile-apps]

Stale CPU Request Comment

The comment still says prod keeps CPU at the chart default 500m, but this values file now sets the request to 900m. Anyone auditing HPA behavior from this file will read the wrong request value and may undo the intended right-size change in a later resource update.

Suggested change

	# the HPA scale on real load. CPU unchanged (500m/2 from the chart default).
	# the HPA scale on real load. CPU request raised to 900m; limit remains 2 cores for bursts.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!