Writing a good incident message
Yann1 min readLire en français
During an outage, your users read every word. What a good message should say, stage by stage.
During an outage, your status page becomes the most read page of your service. A good message does not prevent the outage, but it heads off the wave of “is it down?” messages and keeps your users’ trust.
Three things to say, in every update
- What is not working, in your users’ words: “sign-in fails”, not “502 error on the cluster”.
- What you are doing: you are looking, you found it, you are fixing it.
- When you will give news again, even with nothing new: “next update in 30 minutes”.
One message per stage
In NorminUp, an incident goes through four stages, and each deserves its message:
- Investigating: “Sign-in fails for some users. We are investigating, next update in 30 minutes.”
- Identified: “The outage comes from our database host. We are switching to the backup.”
- Monitoring: “Sign-in works again. We are monitoring for an hour.”
- Resolved: “Everything works normally. Thank you for your patience.”
Saving time with templates
In the middle of an outage, there is no time to look for words. Prepare your messages in advance with incident templates: when the time comes, you only adjust them. Your subscribers get every update by e-mail.