Skip to content

Incidents & Corrections of Error

Post-incident analysis documents capturing what went wrong, why, and what was done to resolve it. These CoE (Correction of Error) documents serve as institutional memory for the cluster.

New incident documents go in docs/operations/incidents/. The three flat files below predate that convention and stay where they are; two of them have nothing to do with DNS or DHCP, which makes relocating them a docs-site change rather than runbook work. Adopt the folder going forward rather than moving history.

Date Incident Severity
2026-08-16 PowerDNS Cutover Walked Forward, Rolled Back, and Re-run Low
2026-08-10 ISP Prefix Re-delegation Left a Dead IPv6 Prefix on Every Client Medium
2026-07-23 WAN Resolver Failure Tore Down Internal DNS Redirect Medium
2026-07-21 Tailscale-on-Router Subnet Route Breakage Medium
2026-02-08 Prometheus HA Split-Brain False Alerts Low