Route mendys-prod / mendys-wp error-log alerts to @claude issues in izzywdev/MendysRobotics

[izzywdev]

Request (from izzywdev/MendysRobotics via delegate-to-fuzeinfra)

@claude — please wire the Grafana crit-log → GitHub issue autofix path (the critlogdetect LogQL rule → crit-alert Cloudflare Worker bridge → grafana-crit-fix.yml) to open @claude issues in izzywdev/MendysRobotics for errors originating in our namespaces.

Root cause

The grafana-crit-fix.yml workflow resolves the owning repo from the alert's namespace label via the namespace annotation fuzeinfra.io/owner-repo. Our namespaces mendys-prod and mendys-wp appear to be missing that annotation, so error logs from mendys-backend/consumers (e.g. Python NameError, Postgres relation "..." does not exist, and a wedged crawler) never opened a @claude fix issue in our repo — they silently fall back to FuzeInfra.

Please

1. Add fuzeinfra.io/owner-repo: izzywdev/MendysRobotics to mendys-prod and mendys-wp — ideally via managedNamespaceMetadata in the Argo Application(s) for these namespaces (mirroring how fuzeinfra's own namespace is annotated in argocd/applications/fuzeinfra-prod.yaml), so it's GitOps-persistent and survives selfHeal.
2. Confirm the automation PAT (GH_TOKEN) used by grafana-crit-fix.yml has issues:write on izzywdev/MendysRobotics.
3. Confirm crit_bridge_token is set in terraform/contabo (so the crit-alert Worker/route + CRIT_BRIDGE_TOKEN secret exist and Grafana's webhook actually reaches the bridge).

Notes

• We've made our backend errors alertable on our side (a mendys_backend_log_errors_total counter + MendysBackendErrorLogs Prometheus rule), but the log→issue→@claude routing is the FuzeInfra half.
• We own only deploy/** and cannot operate the shared cluster — hence this delegation.

Thanks!

[claude]

Claude finished @izzywdev's task in 0s —— View job

---

I'll analyze this and get back to you.

[izzywdev]

@claude — your previous run finished in 0s without doing any work. Please retry: wire the Grafana crit-log → GitHub issue autofix path to open @claude issues in izzywdev/MendysRobotics for mendys-prod/mendys-wp namespaces, per the original request. Check current state of the crit-alert routing config before making changes.

[izzywdev]

⚠️ Claude responder did not run to completion

The responder errored before finishing your request, so no changes were made and no branch was created.

• Reason: Anthropic API rejected preflight (HTTP 400): Your credit balance is too low to access the Anthropic API. Please go to Plans & Billing to upgrade or purchase credits.
• Run log: https://github.com/izzywdev/FuzeInfra/actions/runs/29936932971

This notice is automated and will not retry on its own — the responder stops after one failure and waits. If this looks like an API-side problem (key revoked / out of credits / model or 1M-context beta not entitled / Anthropic outage), an operator should check the ANTHROPIC_API_KEY secret and fund/repair the Anthropic account first, then re-trigger by posting a new comment mentioning the responder.

[izzywdev]

Analysis and delegation:

The mendys-prod and mendys-wp namespaces are managed by MendysRobotics' own Argo Application (not FuzeInfra). The annotation needs to be set via managedNamespaceMetadata in that Argo Application — otherwise selfHeal could revert a FuzeInfra-applied annotation.

Delegated to MendysRobotics#264 to add fuzeinfra.io/owner-repo: izzywdev/MendysRobotics to both namespaces via managedNamespaceMetadata.

On the FuzeInfra side (items 2 and 3 from the ask):

• The GH_TOKEN PAT for grafana-crit-fix.yml would need issues:write on MendysRobotics added — this requires an operator action on the PAT scopes (check Terraform var.github_token scopes; if it's an org-level PAT it likely covers the whole org already)
• crit_bridge_token was rotated recently (see #CRIT token drift incident) and should be current in Terraform/secrets