Production Operations

Day-two work: monitoring, on-call, backfills, deployments and capacity. This path sequences the category's 6 resources by roadmap stage and composite score — work top-down inside each step.

Reliability

1. Runbook Runner 35.6 — Runbook Runner
2. Plan and execute a backfill that doesn't break production 34.15 — Plan and execute a backfill that doesn't break production
3. Production gotchas — silent failures and partial writes 34.15 — Production gotchas — silent failures and partial writes
4. Production Write Confirmation 33.6 — Production Write Confirmation
5. Pipeline failed 28.75 — Pipeline failed
6. Reading CI Failures Across Pipelines, Not One Build at a Time 23.55 — Reading CI Failures Across Pipelines, Not One Build at a Time

• Scope a backfill by date, dependency and blast radius before running anything
• Shadow-write in date batches, smallest first, validating a sample per batch
• Gate the backfill on an explicit range and a rollback path
• Catch silent failures — hidden cost regressions and silently widened freshness SLAs
• Stop partial writes — no appends to recomputable partitions, no unguarded prod writes
• Put the failure class in the alert text so on-call can triage before touching anything

Security-flagged: Runbook Runner, Pipeline failed — review the security posture first. Verification on these records: 6 community. Nothing here is independently audited — the trust scan and your own review are the controls.

Phase 4 of the roadmap. Stage context: Reliability · side-by-side comparison · all learning topics