Spent all day working on those scripts. Oh, first, I did manage to extract logs from datadog via API. Which probably means that next I'll rather do it using scripts, jq, and, maybe, plain Scala, instead of that sophisticated, tricky GUI.
The whole integrated thing does not work together yet, though. Datadog behaves like a bad teacher: you send it a JSON, it says "wrong JSON". What? You send it in the command, -d '{....}', it works. You put it in a file, send it with -d "@$filename", it says the same JSON is wrong.
But ok, now I can reproduce this shit, and I will make it work. Parameterize, to make sure I keep Bermi's code, with a little bit of indirection.In general, I guess I see the problem with scripting. Scripts are written by the people who may be smart, but are clueless regarding programming. They solve one problem, and then another, by just remembering the tricks. Never generalizing. That's why I write scripts with chatGPT. The bot frequently produces something that works, so there. But shit, these inexperienced programmers.
And yes, I did not know what was the problem. The problem is, lack of taste and programming abilities. Like Perl, with Larry who was never a programmer, but a linguist.
Bermi's makefile consists of a bunch of tasks consisting of calling shell scripts. So, what's the point. Who will take makefiles seriously? It's like Roman numerals. There's a funny but stupid discussion on facebook today about a Python code that works in one version, but does not in another. The code relies on the order in which entries of a Set are being enumerated. My attempts to explain that a Set may be enumerable, according to Zorn lemma, but they are not axiomatically and canonically enumerable, were just totally not understood.
Oh, whatever. I'm glad I see working parts (of the clock), tomorrow I'll make it all work together.