Skip to main content
Pre-mortems are AI-generated risk assessments for open pull requests. Before a change ships to production, the pre-mortem agent reviews the PR diff against your service catalog, runbooks, and incident history to identify which parts of your system are most likely to break.

What is a pre-mortem?

A pre-mortem is the opposite of a post-mortem: instead of asking “what went wrong?” after an incident, it asks “what could go wrong?” before a deployment. sureops generates one for each PR that touches services in your environment. Each pre-mortem includes:
  • An overall risk score (0–10)
  • A list of risk factors — specific concerns about the change and the services it touches
  • Affected services — which services in your landscape could be impacted
  • Similar past incidents — historical incidents whose root causes resemble patterns in this diff
  • A recommendation — whether to proceed, proceed with caution, or escalate for further review before merging

The Pre-mortems list

Navigate to Pre-mortems in the sidebar to see all recent pre-mortem assessments. The list is organized with the three highest-risk PRs pinned at the top when no filters are active — these are the changes most likely to cause an incident and deserve attention before they merge. Below the pinned rows, the full list appears newest first.

Filters

Use the filter bar to narrow the list:
  • Repository — filter by the GitHub repo the PR belongs to
  • Status — Pending (assessment in progress), Posted (assessment complete), or Failed (agent could not complete the assessment)
  • Minimum risk score — a slider from 0–10; drag it up to show only high-risk PRs
Filters update the URL so you can bookmark or share a filtered view.

Pre-mortem statuses

Reading a pre-mortem report

Click any row to open the pre-mortem detail page for that PR.

Risk score

The risk score is a 0–10 number computed from the combination of risk factors. Use it as a signal, not a gate:

Risk factors

Each risk factor is a specific concern about the change, such as:
  • “This diff modifies the payment-service database migration path — past incidents show this is a high-blast-radius change”
  • “The changed function handles retry logic; similar retry-loop patterns appeared in 3 past incidents”
  • “Service has no runbook coverage for the error path this change introduces”

Similar past incidents

The agent surfaces historical incidents whose root causes show structural similarity to what this PR is changing. Each match includes the incident title, date, and the pattern that matched.

Recommendation

The final section is the agent’s recommendation: proceed, proceed with caution, or escalate. The recommendation is advisory — merging is always your decision.

High-Risk PRs widget

The three highest-risk open PRs appear in the High-Risk PRs widget on the Command Center dashboard. This surfaces risk at the moment you check in on your environment, not only when you visit the Pre-mortems page.

Triggering a pre-mortem manually

By default, pre-mortems are generated automatically when a PR is opened or updated. If you need to trigger one manually — for example, for a PR that was opened before sureops was connected to your GitHub — use the Run pre-mortem button on the Pre-mortems list page (requires administrator or owner role).

GitHub requirement

Pre-mortems require a connected GitHub integration. When GitHub is not connected, the Pre-mortems page shows an empty state with a link to the integrations settings. See Connect your stack for setup instructions.