Fork me on GitHub
#announcements
<
2020-07-08
>
bruno.bonacci14:07:24

Just released: µ/log v0.3.0 - μ/log is a micro-logging library that logs events and data, not words! It now features publishers for: • Console (stdout) • File • ElasticSearch • Apache Kafka • AWS Kinesis • Slack • OpenZipkin https://github.com/BrunoBonacci/mulog

emmanueloga17:07:02

I'm using these on a small project:

spootnik/unilog {:mvn/version "0.7.24"}
  org.clojure/tools.logging {:mvn/version "1.1.0"}
what would be a good reason to switch to ulog? The things I care about is not having to dig too deep into the java logging conf rabbithole, but I've also been thinking on having structured logging (logging to json, for instance, since my log viewer supports it)

bruno.bonacci18:07:35

µ/log breaks the barriers of traditional logging. It is designed ground up to allow logging of data. Traditional logging system which just provide a JSON formatter don’t truly log events. The interesting data has to be extracted from encoded messages. µ/log provides the facilities to publish events directly from your code. It supports multiple format and multiple systems. The fundamental idea is that events are the essential information that underpins the “3 Pillars of observability”. µ/log provides tooling to prove that. If you just use µ/log in conjunction with Elasticsearch and Zipkin you’ll get more information that many commercial observability products out there. Forget about the existing logging frameworks, just think which information you would like to have when you see your system as a blackbox and instrument your code accordingly. µ/log is just data and you can use it to monitor your system quantitatively, alert on aggregated events, slice and dice the events to narrow down a problem to the source, apply statistical learning and machine learning to extract additional insights. here is an example of some of the question you can ask just by instrumenting 1 part of your system. https://github.com/BrunoBonacci/mulog/issues/25#issuecomment-656062331