Logging with mulog and unstructured messaging like jetty

https://clojurians.slack.com/archives/C010TGGL02X/p1627067453014700

Mixing structured and unstructured logs
E.g including jetty logs with mulog

@ bruno.bonacci is it possible to use mulog to capture the logs from an embedded jetty server that runs my web application (reitit, ring). Or would a more typical java logger to capture the jetty logs and push them to the same log service as mulog

Ah, it seems I would just send both sources to something like elastic search and that would merge the two sources.

Hi @practicalli-john, yes μ/log can do that.
with a single well crafted ring wrapper you can callect enough information to perform a large number of queries (see comment)

this was one of the main design goals of μ/log.

here there is a simple wrapper example

This works, not only on a single application, but if you are in a microservice environment, and each system instruments with a fairly simple wrapper you can get tracing across all your application without the need of expensive tooling and you still get to keep the raw data.

μ/trace captures the relationship of internal calls to other μ/trace as well generating call traces like
https://raw.githubusercontent.com/BrunoBonacci/mulog/master/examples/roads-disruptions/doc/images/disruption-trace.png

There are teams that use it to build complex pipelines to detect quite sophisticate issues in their systems
μ/log -> kafka -> streaming-app -> kafka -> monitoring-dashboard
and it is just the same data that they already use with Elasticsearch or Cloudwatch

(at some point) I hope to make a talk about how teams use μ/log for observability and operation (autoscaling, downscaling, warm-up, etc)

seeing all this in practice makes so much more sense.

Start by reading examples/roads-disruptions/src/com/brunobonacci/disruptions/api.clj around lines 56-69 and the linked μ/log discussion. The issue does not define a concrete change or acceptance condition, so clarify whether Jetty log capture or a documented integration is wanted before implementation; done would require an agreed scope and demonstrated unified logging.