насчет израильских юридических перспектив

Процитируем.

James Madison's declaration in Federalist 47 that "The accumulation of all powers, legislative, executive, and judiciary, in the same hands, whether of one, a few, or many, and whether hereditary, self-appointed, or elective, may justly be pronounced the very definition of tyranny." 

fucking dybr

 I had some fun today.

Yes, I was working on that "two-liner", trying to bypass all those weird behaviors of play json, see https://stackoverflow.com/questions/76147861/scala-play-json-how-to-capture-the-whole-json-source (doable, yes).

And waited for people to approve my prod deployment of that stats update.

Meanwhile people over there were getting pissed off because I was not fixing production problems.

How the fuck can I fix them? "Processor is taking 90% of its capacity" - is not it good? "HAproxy cannot access this sever" - fine, let me look into that server's logs (DynamoDB was the culprit).

But by the end of the day (Karen is on vacation), after Jacob was pissed off that I did not spend (my) night trying to fix the dead Jenkins (0 docs, and nobody was answering my calls), and Davis was kind of confused regarding why I'm not fixing the stuff, I probably convinced Davis to walk through this shit together.

Turned out that HAProxy is running in the same kubernetes whatever (cloud), and we can find its name and ip, and if it decides that it's too busy or cannot receive responses from the servers, we just kill it. And it will resurrect again, this time doing good. Since it's HAProxy, not our code, nobody fucking cares what exactly happened. Just shit happened, kill it.

Then it turned out that it's all written in the workbooks. Except that links in the workbook are too old and don't lead anywhere. And that the aws/kubectl scripts that I had since last year don't work anymore. Because they changed urls. And Swathi even wrote about it in slack. In devops channel. Should I be reading devops channels, figuring out what in my scripts I must change? What was changed, mostly, is eks23 was replaced with eks21. Except that it was not done anywhere. So I could not run my scripts, and instead I had to just run this command and that command, as suggested by Davis. But the result was the same: switch a context, find a pod, kill it. They do it all the time. Manually.

It's 22:25 now, and I'm eventually done with it.

Ok, also, investigated another "user case", and eventually wrote a "spike" regarding some obvious things: "let's log users actions when a profile is being changed", "let's keep the user profile history". Because somebody changed a user's email, and there's no trace. And it was a hack.

Oh, whatever.

Yes, I could retire any moment. But first, what I'll be doing? Walking around the town? And second, stocks just dropped today. Except for GLD that I was thinking about buying, yesterday. Oh, whatever.