log.Info('Always be STRUCTURED LOGGING')

I think of structured logging as an evolution to print statements and even using a debugger. Since with a debugger you have to carefully save your watches and breakpoints if you want to reproduce your session. However with structured logs you would at most adjust your debug level whilst making a query for the state you want to track.

GAME CHANGER for me. Hope it is for you.

[buxt-tech]

Wouldn't it be faster from your applications point of view to push the plaintext logs to a log aggregation tool, which could then parse them at its leisure, and push them to other tools using a structured format?

I get that the purpose here is to speed up incident response time, but I feel like that's already going to be at a human timescale. That leave plenty of time for a log aggregation tool to parse and convert the logs.

I might just be misunderstanding the purpose of this video, though. IN any case, I definitely agree that having structured logs to search through (AND INDEX!) is easier and faster for both humans and machines.

[philipphellmich]

Unfortunately cloud watch is one of the most expensive AWS services. Even if you just use it to stream the logs to kinesis. Some of our AWS bills have a 30+ percentage of cloudwatch logs.

[scottjbarr]

Hey Kai, long time no see!


The "up" tool is really nice. Simple is underrated.