Being On-Call
This page is for engineers who are on call — how to get paged, how to respond, and how to use Regen during an incident.
Getting notified
When an incident is triggered and you are the on-call responder, you will receive:
- A Slack DM from the Regen bot with the incident details and a link
- A message in the incident Slack channel (auto-created for every incident)
- An Adaptive Card in Microsoft Teams (if your team uses Teams)
The notification includes: incident title, severity, triggering alert, and a direct link to the incident.
Responding from Slack
You will be added to the incident channel automatically. From there:
Acknowledge the incident:
/incident ack
or click the Acknowledge button in the card.
This stops the escalation timer and signals to the team that you are working it.
Set yourself as incident commander:
Click Make me Lead in the Slack channel card.
Add a note to the timeline:
Click Add Note or type in the channel — messages are synced to the incident timeline.
Get an AI summary (if OpenAI is configured):
@Fluidify Regen summary
Resolve when done:
/incident resolve
Responding from the UI
Open the incident at https://your-domain.com/incidents/:id
The detail page shows:
- Current status and severity
- Linked alerts with labels
- Full timeline in chronological order
- AI-generated summary (if configured)
- Linked post-mortem (after resolution)
Checking who else is on call
/incident status
Shows the current incident state, commander, and the on-call schedule for context.
Handing off
Before your shift ends, use the Handoff Digest:
- Open the incident in the UI
- Click Generate Handoff Digest
- Regen summarises what happened, current status, and open action items
- Copy it to Slack or email it to the incoming responder
Override your own shift
If you need to hand off early:
- Go to On-Call → Schedules
- Find your schedule and click on your shift
- Click Create Override
- Select the covering person and time window
The covering person is notified via Slack DM.
Tips for effective on-call response
- Acknowledge fast — even if you don't have an answer yet. It stops escalation and tells the team someone is on it.
- Add notes as you go — don't wait until the end. Real-time timeline entries are invaluable for post-mortems and handoffs.
- Set severity correctly — update it if the initial auto-assigned severity is wrong. It affects visibility and escalation.
- Link the root cause alert — if multiple alerts fired, mark the one that caused the others.
- Resolve cleanly — only resolve when the issue is actually fixed, not just when symptoms disappear.