Incidents · Getting started
Incidents
Declare an outage, run it, tell your team and learn from it.
Declare an incident by hand, or from your own systems through the API. Give it a severity and a lead, post updates as it moves from Investigating to Resolved, and your team hears each step by email, Telegram or Slack. Afterwards, keep follow-ups and a write-up on it.
Incidents is for whoever keeps your app running and whoever answers your users while it is broken. It runs in the browser: there is nothing to install, and anything that can make an HTTPS call can open incidents too.
How it works
Each incident has its own reference in the project, like INC-12. It starts as Investigating and moves on to Identified, Monitoring and Resolved, or ends as Declined when it wasn't one after all. Every change is a line on its timeline: who did it, what it was, what it is now.
- Severity: SEV1 Critical, SEV2 Major or SEV3 Minor. Raising it tells your team again.
- A lead: one of the project's people, who runs it. Every notice names them.
- Updates: what you found and what comes next. An update can be an internal note, which never leaves Incidents.
- Notices: by email to the project's people, and to Telegram chats and Slack channels you add, when it is declared, raised, moved on or resolved.
- Follow-ups and a write-up: what to do so it doesn't happen again, each with an owner and a day, and what happened, the cause and what you change.
If the project also has other My Beast products, an incident's timeline shows what they had from the hour before it began: a deploy, a flag turned on, a changed variable, errors and log lines.
Your project's pages are at services.my-beast.com, below the workspace and the project:
| Page | What it shows |
|---|---|
| Overview | The project's incidents, open ones first, with how many began, the median time to resolve and the follow-ups left. At /<workspace>/<project>/incidents. |
| An incident | Where it stands, its timeline, follow-ups, write-up and the notices sent. At /<workspace>/<project>/incidents/incidents/12 for INC-12. |
| Follow-ups | Every follow-up not done yet, the overdue first. |
| Open now | Every open incident in all your projects, under All projects in the sidebar. |
| Settings | The project's keys, and where its notices go, with a test notice. |
What you need
- A project with Incidents added, at id.my-beast.com. The quickstart starts there.
- The people who run incidents in the project. Their role there says what they can do: a Viewer reads, Triage declares and runs incidents, an Admin also manages keys and notices.
- For your own systems to open incidents: a server secret for Incidents. See Open incidents from your systems.
- For Telegram or Slack: a chat your bot can post to, or a Slack incoming webhook.