Skip to content

talkops-ai/talkops-alertmanager

v0.2.0Apache-2.0

Operational skills and MCP execution server for Prometheus Alertmanager triage, notification diagnostics, silence lifecycle management, routing validation, and governance.

auditing-alertmanager-governance

Audits Prometheus Alertmanager alert taxonomy, default-route catch-all leakage, missing runbook/summary annotations, chronic long-firing alerts, control-plane config reload/delivery health, and silence governance compliance across single or multiple backends. Use when asked to run an Alertmanager governance review or scorecard, find unrouted alerts hitting the catch-all route, audit stale zero-match or expiring silences, review denied silence attempts in the audit log, or compare regional Alertmanager posture. Do not use for active incident triage (use triaging-alertmanager-alerts) or creating maintenance silences (use managing-alert-silences).

diagnosing-missed-notifications

Diagnoses why an Alertmanager alert did not page or notify a downstream receiver (Slack, PagerDuty, OpsGenie, email, webhook) and verifies receiver delivery end-to-end. Use when asked why on-call was not paged, why Slack or PagerDuty stopped receiving alerts, to troubleshoot missing alert notifications across Prometheus push, silences, inhibition rules, sibling/parent route shadowing, mute time intervals, grouping timers, or failed receiver deliveries, or to push a synthetic test alert after rotating a receiver webhook or key. Do not use for general alert storm triage (use triaging-alertmanager-alerts) or creating maintenance silences (use managing-alert-silences).

managing-alert-silences

Plans, validates, previews, creates, updates, and expires Prometheus Alertmanager silences for maintenance windows and incident noise suppression. Use when asked to silence or mute an alert, schedule a timezone-aware maintenance window, check silence blast radius, extend or update an active silence, validate silence matchers, or expire a silence early. Do not use for read-only incident triage (use triaging-alertmanager-alerts), static alertmanager.yml mute_time_intervals (use validating-alertmanager-routing), or broad governance audits (use auditing-alertmanager-governance).

triaging-alertmanager-alerts

Triages active and suppressed Prometheus Alertmanager alerts during incidents, alert storms, and on-call shift handovers. Use when asked to summarize firing alerts, run /triage on an alerting incident, prioritize critical incidents, investigate alert storms, inspect Alertmanager dispatch groups, distinguish inhibited vs resolved alerts, or prepare an on-call shift handover report. Do not use for creating silences (use managing-alert-silences), validating routing configs (use validating-alertmanager-routing), or debugging missing receiver notifications (use diagnosing-missed-notifications).

validating-alertmanager-routing

Simulates, lints, designs, and validates Prometheus Alertmanager routing trees, inhibition rules, mute/active time intervals, and receiver configurations. Use when asked who gets paged for an alert, to review or lint alertmanager.yml for shadowed routes or undefined receivers, migrate match/match_re to matchers, design inhibit_rules or calendar time_intervals, or run /setup-monitoring routing checks. Do not use for active incident triage (use triaging-alertmanager-alerts), ad-hoc maintenance silences (use managing-alert-silences), or testing rotated receiver webhooks and missed deliveries (use diagnosing-missed-notifications).