Incident response
Run the incidentwhere the evidence already is.
Alerts, on-call, the timeline and the postmortem sit next to your traces and logs, so nobody spends the incident pasting screenshots into Slack.
INC-482 · checkout latency
FIRING- 14:32:07
alert · checkout p99 above 1s for 2 minutes
Step 1 of 5 · Alert fires
p99 has been over 1 s for two minutes. The alert comes with the slowest trace attached, so you start from a real request.
Press play or click through the steps to follow one incident from the first alert to the postmortem.
Where incident time goes.
Count how much of it is spent on the bug itself.
Today
5 steps · 4 tools- 1PagerDutyacknowledge the page
- 2Slackopen a channel, invite people
- 3Grafanafind the right dashboard
- 4APMlook for a failing trace
- 5Slackpaste screenshots for everyone
With FerrisTrace
2 steps · 1 tool- 1FerrisTracedeclare from the alert with severity, owner and on-call set
- 2FerrisTracethe failing trace is already attached
The first ten minutes go to the bug instead of the channel setup.
From the page to the postmortem.
- Alerts with evidenceSet thresholds on any signal. The alert arrives with the traces that crossed them.
- Severity and ownershipSeverity, owner and on-call are set the moment you open the incident.
- Automatic timelineAlerts, deploys, comments and status changes are added in order as they happen.
- Telemetry in placeThe traces, logs and metrics you attach stay on the incident for whoever joins late.
- Postmortem draftsThe timeline turns into a first draft, so you edit instead of starting from nothing.
- Agent summariesAn AI agent with its own FerrisKey identity can post the first summary.
The rest of FerrisTrace.
Spend your next incident on the fix.
Alerts, on-call and postmortems live in the same tool as your traces.