Fork me on GitHub
#observability
<
2020-09-28
>
plexus08:09:50

regarding "what would you use", I tried to answer that question here: https://lambdaisland.com/blog/2020-06-12-logging-in-clojure-making-sense-of-the-mess obviously there is more then one answer but I've used pedestal.log on a bunch of different projects now and have really come to appreciate its structured logging and integration with logback via slf4j

timo08:09:39

Thanks @plexus. Of course I've read your article:) But regarding metrics and tracing, what did already use? mulog seems to me that it offers all in one, right? just curious about alternatives.

plexus08:09:51

yeah sorry, don't really know much about that 🙂

athomasoriginal13:09:03

@timok Those are good questions. If it helps to contextualize, based on https://clojurians.slack.com/archives/C06MAR553/p1600781157037500 with the author of mulog it seems that mulog is not a replacement for a logging library, but a tool that complements them. So, Plexus does an awesome job of answering the question of which logging libs are available. For mulog though, it answers (based on my understanding) the “sub question” of how can I log specific kinds of data like metrics. e.g. record API latency

đź‘Ť 3
nivekuil19:10:39

whoa, thanks for sharing mulog. That looks like it solves something I've always wanted: separating "what happened here?" from "how should I store it?"

athomasoriginal13:09:15

As another example, one could use a logging library like C.T.L and then also use https://github.com/clj-commons/iapetos. (just another approach). So in this scenario, I see mulog as a decent alternative to iapetos in this scenario