Skip to main content

Work an incident from Slack or the terminal

NovaNova
@nova my app is returning errors since the last deploy

Rather not read it all? Ask Nova this in the Slack thread. It walks you through each step and does the ones it can when you approve.

The investigation's card in Slack or the terminal, answered as in the console. Here it is INC-79, the pods-probe-web crash loop (the full investigation).

Who does what: Dana (on-call engineer, org admin) is on call and answers the card. Bob (developer, org editor) owns pods-probe-web and answers questions in the thread.

You need the Enterprise plan, the Slack skill, Nova in #incidents, and Post results to Slack channel set to #incidents in Nova > Settings > Incidents (Alerts).

1. Dana finds the investigation​

In #incidents, open the alert's thread, or start one: @nova why is pods-probe-web crash-looping?

2. Dana and Bob answer in the thread​

Dana answers the card's buttons; Bob replies in the thread to steer, for example check the database first. What each button does, and when a change can be approved from the thread: Answer from the alert's thread.

3. Dana hands over or stops​

I'll fix it myself takes the fix over; Stop, in the console, ends the investigation (Reject or take it over).

4. Dana answers from the terminal​

Dana runs astroctl nova and types /investigations: the same card, answered the same way (Nova Terminal).

Open zoom view

astroctl nova showing the pods-probe-web investigation: Resolved, the cause with the quoted nginx chown error from App Logs, each step Nova took, the fix with its risk and undo, and pods-probe-web recovered, with keys to scroll, answer, print and go back

astroctl nova showing the pods-probe-web investigation: Resolved, the cause with the quoted nginx chown error from App Logs, each step Nova took, the fix with its risk and undo, and pods-probe-web recovered, with keys to scroll, answer, print and go backastroctl nova showing the pods-probe-web investigation: Resolved, the cause with the quoted nginx chown error from App Logs, each step Nova took, the fix with its risk and undo, and pods-probe-web recovered, with keys to scroll, answer, print and go back

Result: one card, answered from the console, Slack or the terminal; each answer is recorded with who gave it.

Next​

Write the postmortem from the record