dybr

 By the time the day ends, I"m totally exhausted, and hardly remember wtf was going on today.

So, just a list of actions.

1. Read their instructions regarding OKR, and filled some forms. Writing this kind o stuff... did it for the last 40 or something years.

2. Spent too much time trying to push a pretty obvious PR to github - but some of the tests were failing in a weird way.

3. Causes:
  a) An exception thrown inside future is not so trivial to propagate to the top, especially when the top is not on top, but inside a closure;
  b) Found 4 NPEs in AWS libraries; one their own (a variable not initialized - Java!), and another in java library, in which, to do base64 encoding, the author calls some JAXB, which uses some "ModelBuilder", written by Kohsuke Kawaguchi, which pulls some XML, which is not present, so that the "ModelBuilder" just throws an NPE.

4. Cleaned up the code and the Swagger docs - but Swagger is crazy, it counts leading spaces, so there. 

5. I'm on Pager Duty, so there was a huge latency (5 sec instead of 1 sec, and later it was even 20 sec); I found the piece of code that holds this back, but there's no logging there, so go guess.

6. Had a meeting at 20:30, "what should we do to fix stuff"; I posted a poster "do more logging", and everybody was against it.

7. Instead they just guessed that one service (our pardners) was too slow to return the data (as if there's a lot of data and we can't cache). They don't need logs if they can just guess. 

8. Also was told that during PD week I must read emails, because that's where customer support sends us their complaints. Oh shit. I have almost 500 emails in that microsoft box.

That's it for today.