Well Coq has program extraction built in.

Yeah and that's why it's way better than the likes of TLA+.

I would upvote this comment more if I could.

I already refrained from introducing event sourcing to tackle wierd dependecies multiple time just by justaposing the amount of discipline that the team has that lead to the current state vs the discipline that is required to keep the event source solution going.

I don't agree with this cache take. Adding operations to the cache is easy. Taking the django-redis project as an example there are only two levels until you reach redis-py: The cache abstraction and the client abstraction.

The problem with channels is that if you need to touch the ORM you will have to use a sync_to_async call which will block the event loop.

Why don't you post the original broken Python code.

Might as well just implement virtual threads: https://discuss.python.org/t/add-virtual-threads-to-python/9...

What did you try to debug this?

memray and later a custom request wrapper that output python gc statistics. Our main candidates for the leak are: grpc, asgi server itself, psycopg, django channels. All leaked to some degree. Alas, it did not became clear what caused the runaway leak at 30 MB/s. Capturing flamegraphs just before OOM kills would require some more engineering. Granian contained these situations until upgradings later that year made the system more stable to begin with.

Scale articles are too focused on architecture. What about business problems that come with scale. At a certain scale rare events are common many cases cease to be fixable by some random process that involves humans you have to handle a lot more business scenarios with your code.

AWS can make major version upgrades automatically now with less downtime. I think they do the logical replication dance internally.

In my experience PMs often work at a very high level. How things shold work are defined in a incosistent way when we take into account all the user flows, subtelness and restrictions of other systems. So programmers end up doing a significant chunk of the work by refining the specs so that the thing actually makes sense.