gabriel / musehub public
filed by gabriel human · 37 days ago

Docs/testing/post-launch follow-up: remaining Sections 16, 17, 19 items

0 Anchors
— Blast radius
— Churn 30d
0 Proposals

Context

Split out from #164, #165, and #167, all now closed. #166 (Cutover) is also closed — production already cut over on 2026-08-25, retrospectively documented, nothing forward-looking to do there.

This pass ran the actual 7,028-test suite for the first time (previously only confirmed it collects), found and fixed 2 real bugs, confirmed canonical redirects/TLS externally, and confirmed the maintainer-SPOF test now passes (Aaron has independent access via #150).

Section 16 — Documentation remaining

  • System architecture diagram / data-flow diagram
  • Consolidated environment-differences doc (currently scattered)
  • Database migration rollback/forward-repair procedures
  • Database restoration procedure — now writable, backups exist (#157), just not written
  • Secret rotation docs (#177)
  • Certificate/DNS troubleshooting guide
  • Incident response doc
  • Common alarms and remediation — blocked on #174 (no alerts exist yet)
  • Service inventory with owners — blocked on #175
  • Rollback decision tree, contact/escalation info, post-incident doc review process

Section 17 — Testing remaining

  • Fix the 5 test_opengraph_repo_cards.py failures properly — needs a mocked storage backend for these tests rather than requiring live blob storage config to run at all
  • Staging backup/restore testing (production's was done, #157; staging has no backups yet, #177)
  • Production-like integration tests, explicit auth/authz test confirmation
  • Migration testing, failed-migration handling
  • Instance-failure testing, deploy-during-active-traffic testing
  • Load/soak testing, timeout/retry/idempotency behavior under load
  • Tabletop incident exercise
  • Break-glass process test (the mechanism exists per #150, never actually exercised)
  • Clean-workstation access test
  • Dependency-unavailability testing (secret/DB/queue/external API down)
  • Monitoring/alert delivery testing — blocked on #174

Section 19 — Post-launch cadence, ready to schedule

Not blocked on anything technical anymore — just needs to actually get put on a calendar/task tool:

  • Monthly dependency updates
  • Quarterly DR exercise (backups/restore now exist to actually exercise)
  • Quarterly permissions/cost review
  • Annual domain review
◎ Activity
●
gabriel opened this issue 37 days ago
◎
No activity yet. Use the CLI to comment.