Use casesReporting

An uptime incident log that keeps itself current

Mitra keeps your incident log current. Every Friday at 4pm the agent reads the week's Datadog alerts and downtime, then appends a dated entry to a running Notion log. One line says whether things are getting better or worse, and that line goes to Slack.

Runs: Fridays at 4pm
Used by: Engineering · Ops / Chief of Staff

The prompt

Every Friday at 4pm, add this week's Datadog alerts to our Notion incident log. Post one line to #ops comparing it to last week.

Datadog

Notion

Slack

Build this agent

200 credits to start · no credit card

What the agent actually does.

Read the week's alerts
The agent collects every alert and stretch of downtime from the past seven days, with the service and duration.
Group them by service
Alerts are gathered per service, so one flapping check does not read as seven separate incidents.
Append to the log
A new dated section goes on the end of the existing page, and earlier weeks are left exactly as they were.
Add the trend line
One sentence compares this week's alert count and downtime against the weeks before, and says which way they moved.
Drop the line in Slack
The channel gets the one-line version and a link, so nobody opens Notion to know how the week went.

Make it yours.

For the engineer asked how uptime is trending, with nothing to point at but a dashboard that remembers two weeks. Change any of this by saying so — there is nothing to rewire.

Log each environment separately, so staging noise stays out of production history.
Add a monthly run that folds the four weekly entries into one review.
Link any incident with a written postmortem from its entry in the log.

Questions, answered.

How do I keep a running incident log without writing it by hand?

Mitra appends to the same Notion page every week rather than starting a new one, so the log builds into a history you can read back through. The agent reads the alerts and downtime from Datadog and writes the entry in your format. Earlier entries are never rewritten.

Can an agent tell me whether our uptime is getting worse?

Yes, in one line. Mitra compares this week's alert count and downtime with the weeks already in the log, then says which direction they moved. That line sits beside the entry, so a bad week is obvious without a dashboard, and deeper analysis stays your call.