Using ilert AI for post-mortem creation
PreviousUsing ilert AI for efficient incident communicationNextUsing ilert AI for schedule generation
Last updated
Was this helpful?
Last updated
Was this helpful?
As a technical issue transitions from an alert into an incident because it has an effect on business services. Its disruption and impact might make it necessary to revisit what happened and which steps can be taken to prevent a similar issue from reoccurring in the future. Assume following sample incident that has reached its resolved state.
As customers have been notified, the root cause identified and a potential fix in place, the dust and hectic has settled around the incident. Now it is time to sit back and think about what caused the issue in the first place and if/how such things can be prevented in the future.
Usually someone in the team responsible for the services that were impacted by the incident will now start working on gathering all information around the incident and set up a document with all details that is discussed afterwards. This can be quite cumbersome and ilert AI is here to help reduce the time it takes to get started with a proper doc significantly.
If ilert AI is enabled, you will find the option to generate a postmortem in the detail view of the incident.
This will open a popup that offers multiple options to fine tune your postmortem document generation.
By default the generation will take all available context of the incident into account, such as history updates, subscribers, services and involved users / responders.
If an alert is linked to the incident, the alerts details will also be taken into account, as well as its timeline. This can help to significantly improve the generation of technical related action items in the final document. You can also link alerts afterwards, as long as they are not resolved yet.
When an alert is linked that has an attached #warroom Slack channel or MS Teams channel the channel will also appear in the selection options of the popup. Channels that are included will be scanned and parsed by the connected ilert Slack or MS Teams bot and relevant messages will be included automatically in content of the final document.
It is also possible to paste a chat transcript from any other source manually, this script may contain timestamps and usernames in any raw format, ilert AI will automatically parse and handle the content accordingly.
You have three additional fields to provide optional content for fine-tuning of the postmortem document. The content here has a little higher weight in such that during the generation it can help shape action items or root cause. This is helpful if certain action items or root cause information are already known at the time of postmortem generation and you wish to include them.
When you are satisfied with the postmortem input you can click on Generate post mortem to enqueue your postmortem creation. As certain sources are pulled, paginated, read and parsed during the creation process it has been decoupled and can take a little moment.
As soon as your document is ready, the status will transition to created.
You may now click on View generated content to see a simplified markdown rendered version of the generated content:
For our sample alert the generated markdown version looks like this:
At the bottom right you will find the Open option which will contain a link to the raw text file document of your generated postmortem. You will probably want to move this raw version of the document into a shared MS Word or Google Doc for further manual fine-tuning and adjustments before sharing it with colleagues.
As soon as you are ready to share your document with internal or external stakeholders and subscribers you have the option to link your document to an incident.
You may also choose to include your postmortem on all status pages that this incident appears on.
