Monitor post-deploy canary signals and trigger rollback on quality degradation.
npx skills add https://github.com/PiercingXX/xx-stack --skill ops-canary
Please help me install this Agent Skill. Skill: ops-canary Source: https://github.com/PiercingXX/xx-stack/tree/main/runtime/skills/ops-canary Command: npx skills add https://github.com/PiercingXX/xx-stack --skill ops-canary
ops-canary prevents silent production regressions by monitoring real post-deploy signals and enforcing clear rollback triggers when core quality or critical user journeys degrade.
Use ops-canary immediately after deployment to monitor live metrics for 2 hours and automatically initiate rollback if the configured rollback triggers fire.
High-intent search queries and answers about installing and using this skill.
Post-deploy canary monitoring tracks error rate, P95 latency, throughput changes, and critical user journey success using available telemetry. It verifies system health frequently during the first 30 minutes, then continues at a lower cadence for up to 2 hours.
Automatic rollback triggering relies on monitoring live production telemetry and enforcing clear thresholds. When quality signals degrade or critical user journeys fail, the system classifies the outcome as FAIL to initiate an immediate rollback.
Canary deployment outcomes are classified as PASS, FAIL, or AMBIGUOUS. AMBIGUOUS status applies when missing or noisy telemetry prevents deterministic confidence classification, generating a structured report with metric deltas.
Post-deploy health checks apply to release validation scenarios using only the signals that actually exist in your current system. If authentication success is the only available telemetry, it will use that for regression detection.
Deterministic confidence classification confirms genuine regressions from clear telemetry deltas, while transient classification isolates environmental noise. This distinction prevents unnecessary rollbacks caused by temporary capability or environment fluctuations.
g-zenrSet up structured logging, health probes, and optional metrics endpoints.
webrix-aiConfigure incident alerting workflows with Slack and Grafana integrations.
tsuga-devCompute p95 latency per span.name to identify slow trace operations.
aspectrrDetects production issues via console errors, network requests, and performance regressions.
