Incidents
Alerts tell you something is wrong. Incidents let you track what you did about it. An incident is a structured record of an operational event — the timeline, the investigation, the fix, and the lessons learned — all in one place.
What are incidents
An incident is different from an alert. Alerts fire automatically when a condition is met. Incidents are opened deliberately — by escalating an alert, or by creating one manually — when a situation warrants coordinated tracking.
Use incidents when:
- Multiple alerts are firing and they share a root cause
- A problem requires investigation across several nodes or services
- You want a record of what happened, what was tried, and how it was resolved
- Multiple people need to coordinate on a response
Not every alert needs an incident. A disk-full alert on a dev VM that you fix in five minutes is probably just an alert. A storage controller failure that takes down three production services is an incident.
Incident workflow
Opening an incident:
- From the alerts view, escalate an active alert instance to create an incident pre-populated with the alert context.
- Or create an incident manually from the Incidents page with a title and initial scope.
Scoping impact:
Identify the affected sites, nodes, and services. LabTether's topology and service views help you trace the blast radius — which downstream services depend on the failing component? Which upstream systems might be contributing?
Building the timeline:
Add entries as you investigate and act. The timeline is your operational logbook:
- What you observed
- What diagnostics you ran
- What changes you made
- What the results were
Each entry is timestamped, so the timeline doubles as an audit trail.
Coordinating remediation:
Assign action steps, execute fixes, and monitor recovery through the incident view. Link to relevant nodes, alerts, and services so everything is in context.
Resolution
When the situation is stabilized:
- Verify affected systems have recovered — check node status, service health, and alert state.
- Add a resolution summary: root cause, fix applied, and any prevention measures for next time.
- Close the incident.
The closed incident remains in the system as a permanent record. Come back to it during post-mortems or when a similar failure mode appears again.
Audit trail
Every incident captures a complete timeline — who did what, when, and what happened next. This is valuable beyond the immediate response:
- Post-mortems — review the timeline to identify where response time could improve
- Pattern detection — recurring incident types suggest systemic issues worth addressing with automation or infrastructure changes
- Compliance — for small teams managing production infrastructure, the incident timeline provides the operational record auditors want to see
- Onboarding — new team members can read past incidents to understand the environment's failure modes and how the team responds
Tip: Even if you are the only operator, get in the habit of recording incidents for non-trivial outages. Future-you will appreciate having a record of what happened and why, especially when the same symptoms show up six months later.