Skip to the content.

The risk register

Where accepted risks live once a human has agreed they are real.

Premortem is not primarily a register — it is what fills one. But a risk nobody can find later is a risk nobody acts on, so accepted risks get written somewhere durable.


The flow

Propose → decide → write. Nothing is written until the third step.

  1. The assessment proposes risks in the panel. They exist only there
  2. A person marks each accepted, mitigating, or dismissed
  3. Write accepted writes them out

Two checkboxes control where:

Use both. They answer different questions: the description answers “what should I worry about in this ticket”, the register answers “what are we carrying across the project”.


Setting up the Confluence side

⚙ → Apps → Premortem settings → Confluence space(s) for the risk register.

Premortem creates a page called Risk register BT (the project key) in the mapped space if it does not exist yet.

What a row contains

Column  
Task linked back to the Jira issue
Risk what could go wrong
Category security, data integrity, operational, and so on
Priority the action priority and the three ratings behind it, e.g. high · S5 · O3 · D4
Mitigation (prevent) what stops it happening
If it happens what to do when it happens anyway
Trigger the signal that says it is happening now
Status proposed, accepted, mitigating, dismissed

The last three columns are the ones that make a register usable in an incident rather than only in a planning meeting.

The page is set to full width, and each column gets a width of its own — eight equal columns of prose wrap to one word per line. Registers created by older versions are upgraded in place the next time an issue writes to them.

Newest first

New rows go directly under the header, not at the bottom. A register that grows downwards buries the issue somebody just worked on under every issue that came before it, and the page is read far more often than it is written. Writing an issue again lifts its rows back to the top.

Writing again is safe

Rows for an issue are replaced, not appended. Write the same issue five times and the register holds one current set, not five copies. The same is true of the description section — it is rewritten in place, not stacked.


Keeping it from rotting

A register nobody trims stops being read, and then stops being written.

⚙ → Apps → Premortem settings → Register maintenance scans every page titled Risk register * across the site, removes rows whose Jira issue no longer exists, and saves a page only when something actually changed.

It also runs once a week on its own, so in practice you rarely need the button. It is there for when you have just deleted a batch of issues and want the register correct now rather than tomorrow.