Apr. 11th, 2023
Ok, I was pretty nervous, why t.f. do not the keys I post get into the system on prod? That's why I added logging and tried to get an approval to deploy this stuff to staging.
Of course my innovative way of logging (setting time limit) was not understood, and they, traditionally, asked me to use
Oh, whatever. It's really not in my interests to push the boundaries of vision in the conservative people that are basically responsible for the functioning of the system. Should I replace them? No way.
But how will I figure out wtf is going on?
The problem is, we don't have any tests that contact staging or even prod and check the behavior. So I'll need to have some. Meanwhile, I was just running scripts.
And I found that, hmm, the data gets into the database. (I have
OMFG. It's cache. I have cache. I intentionally used guava cache (no shit, it's something that I, together with Bob Lee, aka
So, everything works, and I was just paranoid. Pinging cache to refresh would not help, because there's more than one instance anyway. And it was intentional: do not rush, it'll settle, we'll have eventual consistency.
Fck. Of course, the mood has changed.
Then I went for a walk, then we had a meeting with Jorge, to figure out wtf is happening with using ZIO 2.0 in this project. Ha. It was Scala 2.11 - who, in year 2023, uses Scala 2.11? Conservatives only. So I decided to switch to Scala 2.12 (why not 2.13?) - and that's where I'm currently stuck.
Then we had a 1-1 with Karen, and talked about logs, and, since I got on Kibana through all the logs produced by this app,
That was all for fun, because I did not plan to take the next project today. Tomorrow, tomorrow. Mañana. Morgen, morgen, nur nicht heute. Утро начинается с рассвета. Manhã a bonita manhã.
At this moment I had almost closed my work computer.
But not exactly. There was one more meeting, where Davis suggested to discuss a new endpoint. What's there to discuss, the endpoint exists already, and he wanted to give it an external access, and was asking whether it's secure. It sure is.
Of course my innovative way of logging (setting time limit) was not understood, and they, traditionally, asked me to use
logger.debug(), and update the logback.xml to enable this logging. See, the operation happens several times a year maybe, but we are saving on disk space. And of course if you enable debugging in the config file, who is going to disable it? I see comments in those files, like "This is here very temporarily, will remove it soon" - from over 2 years ago.Oh, whatever. It's really not in my interests to push the boundaries of vision in the conservative people that are basically responsible for the functioning of the system. Should I replace them? No way.
But how will I figure out wtf is going on?
The problem is, we don't have any tests that contact staging or even prod and check the behavior. So I'll need to have some. Meanwhile, I was just running scripts.
And I found that, hmm, the data gets into the database. (I have
psql open.) But the server was not showing it. And then, after I took a break, it did.OMFG. It's cache. I have cache. I intentionally used guava cache (no shit, it's something that I, together with Bob Lee, aka
crazybob wrote when I was in Google... then, of course, there were lots of refactorings.So, everything works, and I was just paranoid. Pinging cache to refresh would not help, because there's more than one instance anyway. And it was intentional: do not rush, it'll settle, we'll have eventual consistency.
Fck. Of course, the mood has changed.
Then I went for a walk, then we had a meeting with Jorge, to figure out wtf is happening with using ZIO 2.0 in this project. Ha. It was Scala 2.11 - who, in year 2023, uses Scala 2.11? Conservatives only. So I decided to switch to Scala 2.12 (why not 2.13?) - and that's where I'm currently stuck.
Then we had a 1-1 with Karen, and talked about logs, and, since I got on Kibana through all the logs produced by this app,
oauth, I know what this code logs - mostly spam. Karen suggested me to clean the logging, I suggested to add memory for logs. And went ahead with looking into where all this shit is coming from. It's not from our code. It's from k8s. Basically, none of our business. Just filter shit out. Ok, found some weird error messages, and found that I could improve the meaning, so I updated just one little line, to make the log message more meaningful.That was all for fun, because I did not plan to take the next project today. Tomorrow, tomorrow. Mañana. Morgen, morgen, nur nicht heute. Утро начинается с рассвета. Manhã a bonita manhã.
At this moment I had almost closed my work computer.
But not exactly. There was one more meeting, where Davis suggested to discuss a new endpoint. What's there to discuss, the endpoint exists already, and he wanted to give it an external access, and was asking whether it's secure. It sure is.