Repository navigation
fix: tighten openclaw and claude-remote memory limits to 2x request - #532
Merged
Merged
Conversation
admins/labs/ovk openclaw and labs/claude-remote carried memory limits 2.5x-4x their requests, a ratio built up reactively during past CPU incidents rather than tuned for memory. Tenant quota usage data shows this headroom sitting unused. Trim to base-service's own 2x default ratio to reclaim quota-ceiling headroom for onboarding more tenants, and apply the same ratio in the openclaw onboarding template so new tenants don't inherit the old ratio. CPU limits are untouched.
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
|
Claude encountered an error after 0s —— View job I'll analyze this and get back to you. |
|
Claude finished @mashkovd's task in 50s —— View job ✅ No P1/P2 findings, good to merge All five memory limit changes are correct and consistent:
All ratios match base-service's default 2× ratio. Comments clearly explain the audit rationale. YAML formatting follows conventions. Commit message is conventional. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
admins/openclaw,labs/openclaw,ovk/openclaw, andlabs/claude-remotefrom 2.5x-4x their memory requests down to base-service's own 2x default ratio.argo-workflows/service-templates/openclaw/values.yaml.tpl) so new tenants don't inherit the old, wider ratio.s3-syncsidecar are untouched (separate CPU-overcommit follow-up, out of scope here).Why
Part of the 2026-07 platform audit follow-up (
docs/audit/PLATFORM_ASSESSMENT_2026-07.md): by memory limits, the cluster only has headroom for ~1 morenfc-sized tenant even though request headroom covers 8-11. These four services carry limit:request ratios far above the chart default, a pattern built up reactively during past CPU-throttling incidents rather than tuned for memory — tenant quota usage data (mctl_get_resource_usage) shows this headroom sitting unused. This is step 1 of a 2-step plan; step 2 will right-size the affected tenants'quotas.limits.memoryceilings to actually return the freed memory to the shared cluster pool.Test plan
helm templaterendered for all four edited values files, confirmed the new limit value appears on the correct (main) containeradmins-openclaw,ovk-openclaw,labs-claude-remote(labs-openclawis currently scaled to 0)@claude reviewand CI green before merge