More talks in the program:
When it comes to logging, running a container based microservice infrastructure offers you some challenges – to put it mildly. In a volatile environment, you can’t just debug a user session by digging through a single server’s log files. Pretty soon you realise that you need to centralise collection, processing and storage of your logs.
Unfortunately, a centralised logging system does not play well with a “shared nothing” architecture approach. If the teams take “share nothing” literally, each team would have to run their own logging stack. Let’s be honest: Most teams are lacking time, knowledge (or both) to do so and the approach is bound to break here.
We’d like to show you in this session, how Shopping24 managed to find a way: We defined a common interface and provided an extensible default log processing implementation. Only a small layer of governance is needed to ensure that the teams have full control over their logs while still being able to participate in version updates and new features easily.