Incident Management Skill

Use this skill as the gateway for incident triage, SLO monitoring, and notification verification. It orchestrates the full triage workflow - from detection through resolution - and cross-references cx-alerts for deep alert management and cx-telemetry-querying for root cause investigation.

CLI Commands

Key flags:

• cx incidents list supports --status (TRIGGERED, ACKNOWLEDGED, RESOLVED), --severity (CRITICAL, WARNING, INFO), --assignee
• All commands support -o json for structured output and -p <profile> for profile selection
• cx slos create/update use --from-file <path> (or - for stdin)

Incident Triage Workflow

Step 1: Check Active Incidents

cx incidents list -o json
cx incidents list --status TRIGGERED -o json
cx incidents list --severity CRITICAL -o json

Get an overview of what's happening. Filter by severity for immediate priorities:

cx incidents list -o json | jq '[.[] | select(.severity == "CRITICAL") | {id, name, status, severity, started_at}]'

Step 2: Get Incident Details

cx incidents get <incident-id> -o json
cx incidents events --incident-id <incident-id> -o json

Review the incident timeline and related events to understand scope and progression.

Step 3: Check Related Alerts

cx alerts list -o json

Find which alerts are currently firing. For deep alert inspection, switch to the cx-alerts skill.

cx alerts list -o json | jq '[.[] | select(.is_active == true) | {id, name, severity, last_triggered}]'

Step 4: Review SLO Status

cx slos list -o json
cx slos get <slo-id> -o json

Check if SLOs are breaching or error budgets are burned:

cx slos list -o json | jq '[.[] | {name, status, remaining_budget_percentage}]'

Step 5: Verify Notifications

cx notifications connectors list -o json
cx notifications routers list -o json
cx notifications presets list -o json

Confirm the right people were notified through the correct channels.

Step 6: Pivot to Root Cause

Switch to the cx-telemetry-querying skill to investigate the underlying cause using logs, traces, and metrics.

Incident Actions

Acknowledge

cx incidents acknowledge <incident-id>
cx incidents acknowledge <id1> <id2> <id3>

Resolve

cx incidents resolve <incident-id>
cx incidents resolve <id1> <id2> <id3>

Assign

cx incidents assign <incident-id> --user-id <user-id>

Close

cx incidents close <incident-id>

SLO Management

Creating SLOs

Template from an existing SLO:

cx slos get <existing-slo-id> -o json > slo-template.json
# Edit slo-template.json with new service/threshold
cx slos create --from-file slo-template.json

Monitoring SLO Health

# All SLOs with their status
cx slos list -o json | jq '[.[] | {name, status, target_percentage, remaining_budget}]'

# SLOs that are breaching
cx slos list -o json | jq '[.[] | select(.status != "OK")]'

Notification Debugging

When notifications aren't reaching the right people:

1. Check Connectors

cx notifications connectors list -o json | jq '[.[] | {id, name, type}]'

Verify the expected channels (Slack, PagerDuty, email) exist and are configured.

2. Check Routers

cx notifications routers list -o json | jq '[.[] | {id, name, entity_type}]'

Verify routing rules map the right alert types to the right connectors.

3. Test Notification Delivery

cx notifications test connector --from-file test-connector.json
cx notifications test destination --from-file test-destination.json
cx notifications test preset --from-file test-preset.json
cx notifications test routing-condition --from-file test-condition.json

Incident Aggregations

Get a high-level view of incident patterns:

cx incidents aggregations -o json

Use this to understand incident frequency, MTTR trends, and severity distribution.

Key Principles

• Triage before deep-dive - check incidents, alerts, and SLOs before querying telemetry data
• Check SLO burn rate, not just status - a slowly burning SLO needs attention before it breaches
• Verify notification chain end-to-end - connector exists → router maps correctly → test delivery works
• Cross-reference with telemetry - use cx-telemetry-querying skill for root cause after triage
• Acknowledge promptly - acknowledge incidents to signal ownership and stop re-notifications
• Use incident events for timeline - cx incidents events shows the full incident lifecycle

Related Skills

• cx-alerts - deep alert management: creating, updating, and inspecting alert definitions
• cx-telemetry-querying - root cause investigation using logs, metrics, traces, and RUM
• cx-observability-setup - configure notification channels and routing for alerts