PagerDuty StrawberryThe 03:12 page, worked and written up by nine
A companion pulls the incident, its full timeline, the service and who is on call, then acknowledges, notes, snoozes or resolves it as a named user when you say so.
The page fires at 03:12. Someone acknowledges it, restarts something and goes back to bed. At 09:00 nobody can reconstruct what happened without scrolling the timeline, finding the note, and cross-referencing the escalation policy to work out who was even on call.
A companion assembles it instead. It pulls the incident’s complete log (triggered, acknowledged, notes, escalations, resolved, with the channel each entry came through) plus the two related incidents on the same service, and writes the timeline section of the post-mortem.
| PagerDuty’s app and automation | A Strawberry companion | |
|---|---|---|
| Grouping and suppressing noisy alerts | Event intelligence, before an incident exists | Works the incident that intelligence created |
| Acknowledging and resolving from a phone | Yes, in seconds | Yes, attributed to a named user, confirmed first |
| Reconstructing a full incident timeline | Scroll the incident log | Pulled as structured entries and written up |
| Answering “who is on call for payments this weekend” | Yes, across schedule views | One question, filtered by schedule, policy and time window |
| Turning last month’s incidents into a digest | Build a report | A routine: which services paged most, which took longest |
Who is actually on call right now?
One question, one answer: current and upcoming shifts, filtered by schedule or escalation policy, over a time window you name. Four schedules and layered policies make this a click-hunt in the interface.
The awkward version works the same way. Which services does this team own, and which of them paged anybody last week. It lists the services and teams, filters incidents to that scope, and returns the count with the incidents underneath it.
Work the incident from wherever you are
Acknowledging, snoozing, noting and resolving are all tools, and each acts as the email address of a real PagerDuty user, so every entry in the timeline carries a named person. Snooze suppresses re-notification for a fixed window, which is the honest version of "I am on it, give me an hour".
These change a system other people trust, so each one is confirmed rather than assumed.
- Acknowledge, snooze for a window, add a note, or resolve, attributed to a named user.
- Update an incident: reassign, retitle, change urgency or priority, attach a bridge.
- Create an incident against a service when something needs a formal record.
- Read the full log: every state change and note, with the channel it came through.
Write the post-mortem timeline
Triggered at 03:12, acknowledged at 03:14 via the mobile app, a note at 03:31, escalated at 03:47, resolved at 04:02. Structured entries rather than a scrolling panel.
A companion turns that into the timeline section of the write-up and leaves the analysis to you.
What does a monthly incident digest look like?
Which services paged most, which incidents took longest to acknowledge, and how often the same alert came back. It filters incidents by status, service, team, user, urgency and date window, counts them, and lists the incidents under the count.
Run monthly as a routine, that is research on your own operational history, the kind promised after every bad week. For the error underneath the page, the Sentry tools sit on the other side of the same incident.
Experience Strawberry for free
DownloadTrusted by fast-growing companies worldwide
Frequently asked questions
Strawberry is free to download and includes AI credits to start. Paid plans begin at $20/month. See pricing. · Reviewed · Canonical facts for AI agents