Register physical infrastructure alert owner routes - #512
Conversation
|
Hey @allyblockcast[bot]! Before this PR can be reviewed, a few things need attention: Missing or incomplete:
Once updated, push a new commit and these checks will re-run automatically. — commitperclip |
|
Hey @allyblockcast[bot]! Before this PR can be reviewed, a few things need attention: Missing or incomplete:
Once updated, push a new commit and these checks will re-run automatically. — commitperclip |
|
Hey @allyblockcast[bot]! Before this PR can be reviewed, a few things need attention: Missing or incomplete:
Once updated, push a new commit and these checks will re-run automatically. — commitperclip |
|
Hey @allyblockcast[bot]! Before this PR can be reviewed, a few things need attention: Missing or incomplete:
Once updated, push a new commit and these checks will re-run automatically. — commitperclip |
|
@ally please review the Alertmanager owner-map change for BLO-12202, focusing on whether the shipped |
Thinking Path
Linked Issues or Issue Description
No GitHub issue exists. Paperclip task: /BLO/issues/BLO-12202.
Subsystem affected
packages/plugins — plugin system
Problem or motivation
Physical infrastructure alerts delivered by Alertmanager need explicit Paperclip owner-map class routes so they create or update owned issues instead of defaulting to unassigned. The new alert families are Proxmox, Ceph, BMC/iDRAC, and disk/RAID/SMART health.
Proposed solution
Register
physical_infra_proxmox,physical_infra_ceph,physical_infra_bmc, andphysical_infra_diskin the bundled Alertmanager pluginDEFAULT_OWNER_MAP.class, add resolver coverage for each class, and document the expected owner/escalation policy.Alternatives considered
Relying only on runtime plugin settings would work after manual configuration but would still leave fresh deploys and reinstalls without durable routing defaults. Routing by
alertnamewould be more brittle because each telemetry area can contain multiple alert rules.Roadmap alignment
This is operational alert routing for the Paperclip plugin system, not roadmap-level core feature work.
Additional context
Acceptance requires each synthetic physical-infra alert class to create or update a Paperclip issue under the intended owner/queue while existing alert classes continue routing unchanged.
What Changed
physical_infra_proxmox,physical_infra_ceph,physical_infra_bmc, andphysical_infra_diskinDEFAULT_OWNER_MAP.class.Verification
Blockcast/paperclip.pnpm --filter paperclip-plugin-alertmanager testcoverssrc/__tests__/owner-resolver.test.ts.pnpm --filter paperclip-plugin-alertmanager typecheck.pnpm --filter paperclip-plugin-alertmanager build.Risks
support@blockcast.net) rather than a more specific physical-infra queue. This matches existing shipped operational routes and avoids unassigned alerts on fresh deploys; narrower routing can be applied by instanceownerMapoverride without code changes.labels.classto one of the registered class names.Model Used
OpenAI GPT-5.5, exact model ID
openai/gpt-5.5, tool-using coding/review agent.Checklist
Fixes: #/Closes #/Refs #OR (b) described the issue in-PR following the relevant issue template