Rolling your own zero-downtime deployments is as about as a good idea as rolling your own security... it's not a good idea.

I run our small system on a single EC2 instance with K3s. It runs a half-dozen or so services and does it quite well. I don't think it is particularly complex or over engineered. I like how easy it is to maintain the configuration in a helm package and quickly deploy it in different environments.

There is a learning curve for doing K8s well, but that's true for any non-trivial system.

I tried using it in a small k8s cluster on digital ocean. Initial installation using the recommended helm package was easy enough. However, it only saved a very short period of log data. I spent a fair amount of time searching docs and the web about how to increase the storage, with no luck. Such an obvious and common need should not be so difficult to configure. You should not have to deep dive, reverse engineer, and read source code in order to solve such simple problems.

This reminds me of Lance Armstrong. Prior to his outing, I believed Lance Armstrong was innocent of doping. But many serious cyclists told me there was no way he was not doping, and they were right. Expertise is often useful.

Right now, I'm running 50+ miles a week training for a marathon. That is really hard and a major commitment. I won't set any records, but I have a good shot at placing in my age group. I don't run to win anything or achieve a certain place, but I would be somewhat pissed off if some idiot like this cheated and beat me.

I enjoy most of the time that I spend running, but to get to this level, you have to go through a pretty difficult level of training which makes all of your life a lot more stressful during the periods of intense training. Right now, I'm staying in a job that I don't like because the schedule is very flexible and allows me to spend around 15 hours a week training.

Apple recently required that anyone using 3rd party authentication (google, fb, etc) in their iOS app, must also provide support for Apple ID signin. That is not an unreasonable request. I was able to implement the signin without too much difficulty. However, when it came to doing a server-side validation of the signin via the Apple API; it was an absolute POS blackbox.

Their documentation and server responses provided no indication of why the validation failed. I searched the docs and forums; it was clear that a ton of other people had the same problem with absolutely no clues into resolving the issue. So I just gave up on it. Total waste of time.

REST is simple, simple is good. It's not perfect for everything, but it has served my purposes well for many years.