It's called a 'post mortem' in the medical world because the patient died. It involves a high level of inspection into what went wrong and what can be learned to prevent it. I assume the term was adopted from there into project management.

It might make a little more sense in the world of shipping software in retail boxes, where products/projects had a 'done' date. The project is dead, what contributed to it's demise? Or you might generalize death into failure, and that's why we use the term instead of post-incident.

Not all post-mortems are for failures/dead products which might add even more confusion. For instance Gamasutra runs a game development post-mortem section where developers of popular (and unpopular) games can hop on and describe difficult situations they've encountered, how they did what they did, why they did that thing, etc.

It's really fun to read.

http://www.gamasutra.com/features/postmortem/

edit: Wow, there hasn't been one since 2014. I wonder why this died out. There's 10 pages of them since 2007..

Ironic. Apparently we need a postmortem postmortem

The earliest I know of: https://books.google.com/books?id=mskkmVkpIUcC&lpg=PP1&pg=PP...

I've seen the act of analyzing project failures by this name in software engineering/management ever since.

Maybe it was a bit more relevant in the days of large waterfall software project management, where failure often meant the end of a project with no product launched. Sometimes after a "death march": https://en.wikipedia.org/wiki/Death_march_(project_managemen...

But it does seem natural to me that it has been carried over to current days, and applied to analyzing failures in the context most relevant to modern software development.

Never stopped to think about the weirdness of the term's application. It feels so natural to me, I'd suggest we all start calling the fixed service as "resurrected". As in: "Google Cloud Global Load Balancers were resurrected at 13:19." (-:

That seems to make sense indeed. It does cover the load in terms of a thorough investigation so I get why it stuck as a more powerful term than a mere incident report.

The term 'post-mortem' doesn't appear anywhere on that page. It was added by the OP in the title. In reality it should be called an incident report.

If we want to be technical, a post-mortem in the tech world is commonly used to outline failures that occurred during a normal event (i.e. a software release) not a random production issue.

Well the outage itself died. Hopefully. Otherwise it would just be an incident status update.