preface

In the consulting experience, I found that some software projects often have online accidents. After the occurrence of online accidents, I will fix the problem at the first time, and the second time, I will hold the software accountable.

This is a very interesting phenomenon, usually in some traditional industry or government background teams, when an online accident happens, they will start the accountability process, find the person in charge of the accident, and punish him accordingly.

As programmers, we all know that the world of code is impossible without errors. Accountability is a big part of how team members are afraid to write code, to go online, to touch everything in the online environment, which ultimately leads to a loss of team productivity.

So what’s the right thing to do?

“Blameless Postmortem report” is a Blameless Postmortem report.

What is a no-fault coroner’s report?

A no-fault coroner’s report is a written record of an online incident that describes:

  • The impact of accidents on this front line.
  • Action taken to mitigate or resolve an accident.
  • The root cause of the accident.
  • Follow-up actions taken to prevent a recurrence of the accident.

No fault autopsy report is a literal translation of the name, and if that sounds too gory, you might call it a no-fault reflection report, or a no-fault accident report, or a no-fault post-mortem report. But more people are used to affectionately calling it a coroner’s report.

No fault is emphasized because it prevents people from blaming each other or hiding their faults for fear of being held accountable when writing a report.

Why do we need a no-fault autopsy report?

The goal of the autopsy report is to understand all the root causes of the accident, document the accident for future reference, and develop effective preventive measures to reduce the likelihood of the accident happening again.

For autopsy reports to be effective in reducing repeat incidents, the summary process must motivate teams to identify root causes and repair them.

At the same time, paying attention to the process and ensuring that it is effective requires commitment at all levels of the organization. There can be no accountability for one person on the team.

When do you need a no-fault autopsy report?

Online accidents have a degree of severity or impact rating, so we usually only write autopsy reports for higher level incidents.

We usually start writing the autopsy report at two points in time:

  • Repair Accident Period
  • After repairing the accident

Who completes the autopsy report?

The delivery team to which the accident occurred is jointly responsible for completing the autopsy report.

But you need to choose an owner who is primarily responsible for writing the report, and this owner needs to ensure that the following two things happen:

  1. Assign different people to complete all kinds of accident investigation work, and finally summarize the results to the owner.
  2. Ensure that the improvement actions in the report are scheduled in subsequent iterations according to their urgency.

How do I keep track of actions in my report?

This question is actually the second one immediately above.

Actions in reports are generally divided into two categories:

  1. Root cause improvement
  2. Non-root cause improvement

For each Action in the report:

  • Cards should be built in the corresponding team backlog and scheduled into the appropriate iteration based on priority.
  • The Owner is responsible for the completion of the tracking card. And record it in the report.

Coroner’s report meeting

There are two kinds of meetings related to the autopsy report.

  1. One is to discuss the root cause of the accident before writing the report.
  2. One is to share the report content with the team after the report is completed to learn and grow.

Whatever the meeting, remember that this is not a criticism meeting where you can point fingers at anyone.

The guidelines here are similar to Retro.

In practice, I found that most teams only held the first meeting, and in the process of writing the report, everyone basically learned and understood the root cause of the report. So most teams don’t have a second meeting.

However, many companies will hold another kind of meeting, which is a report meeting for the leaders after the report is finished. This meeting is optional according to the different policies of different companies.

The report template

A complete report template is shown below.

Reports can be in the form of tables or documents. With a template, the writer of the report can follow the template and fill in the contents.

As for how to identify root causes, Atlassian provides an analysis method called 5 Whys. For details, please refer to the following:

https://www.atlassian.com/tea…

conclusion

The autopsy report is intended to be continuously improved during software development and project delivery, to be documented, and to be part of the knowledge sink.

So it can take the form of the actual team. I’ve seen teams write them in sheets, like the template above, or on cards, or on whiteboards and then take pictures.

For the no-fault autopsy report, if you have any questions in practice, please come to me to discuss.


The resources

https://www.atlassian.com/inc…

https://www.atlassian.com/inc…