I'm really frustrated with our postmortem process. Every week, we write long reports about incidents only for the same issues to happen again a few months later. I just spent two hours detailing an outage in our authentication service, and it's the same root cause as an incident from January. Everyone acknowledges our technical debt, but nothing gets done because we're all too busy dealing with new issues. Leadership keeps asking why we haven't learned from these incidents, but we do learn; we just can't find the time to actually implement fixes. Our retrospectives have turned into lengthy compliance documents without any real impact—half of the action items from last quarter are still sitting in our backlog. I'm looking for tools that can help make our postmortems more useful and prevent them from becoming just another graveyard for Jira tickets. Has anyone found a solution that actually addresses these problems?
5 Answers
One approach is to document previous incidents and highlight management’s lack of prioritization for fixing the recurring issues as a primary action item. When another incident happens, you can point back to your documentation, show that it's already been acknowledged, and then just refer to it instead of reinventing the wheel. This can help avoid repetitive digging into root causes since it’s basically accepted that it’s an ongoing issue.
That's a smart way to repurpose old info! It can definitely save time and keep everyone on the same page.
Make sure to actually do the work to keep moving forward. Gaining some consistency can really help improve your team's efficiency over time and keep those incidents from piling up.
Stop pretending the current process works—be vocal about needing concrete plans for implementing lessons learned from past incidents. It’s crucial to hold your leadership accountable for resource allocation and make them own their decisions regarding team priorities. And if they’re not responding to the reality, it may be time for a self-reflection on your career options!
After a few incidents, try pushing for a formal Problem Resolution process. You might also want to implement an error budget, which can give you the breathing room needed to address ongoing issues without it feeling impossible.
We found that creating a Jira ticket from each postmortem as an action point helped us track repeat incidents better. It gives you documented evidence that there’s a backlog of unresolved issues, and when managers ask why something isn’t fixed, you have a clear response.
Totally agree! Just link your past action items to the current incidents and make sure everyone knows this is the way forward.