OpenShift includes EFK and I'm looking into using that or ELK, but there are so many options (found a link with 50+ options) that I'm trying to narrow down. Ideally something free/cheap. What is the best logging and metrics reporting tool?

We are running a handful of microservices within OpenShift.

Some simple use cases of what we are looking to do:

  • As an admin, I need a report showing how many times a service was called in a given time frame.
  • As an admin, I need a report showing how many times a service was called successfully in a given time frame.
  • As an admin, I need a report showing the average time a service takes to process requests.
  • As an admin, I need a report showing the lifecyle of a specific message (UUID). The lifecycle will show each service called.

The question hasn't had an answer in seven months so I will promote my upvoted comment to an answer:

Finding the lifecycle of a specific message across services is known as "distributed tracing". Distributed tracing requires more than just capturing logs. OpenShift uses Kubernetes which is part of cncf.io/projects which governs opentracing.io and has one tracing solution jaegertracing.io. As Jaeger is a sister project to Kubernetes you are likely to find it widely supported. Here is an openshift commons video about it.

I should mention that the question is likely to gather a lot of answers of the type "the only tool I have used is the best tool ever!". Experts will avoid answering such questions. On many sister StackExchange sites, often questions that get such answers get put on hold by moderators as been unhelpful.

I recommend that you don't look at new tools before fully exploring the tools that are near to hand or "most obviously stable and well supported". As EFK (sometimes more amusingly know as FEK as it sounds naughty) ships with OpenShift its worth really digging into what it can do. Then ask questions about how to do the things that it appears it cannot do. That is the style of question that helps everyone and that experts will gladly answer and you may then find that the most stable and widely support tool meets your needs. That is the best outcomes as it is going to be the easiest to get running, the least likely to break on upgrades, and the least likely to itself go down causing you late nights fixing it.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.