Why should use Prometheus-Grafana and Alert Manager over ELK Stack?

[Varun Sharma]

So I am an Application Developer and have worked with ELK stack for quite a while now. For me, Prometheus seems quite appealing for some of our Use-Cases but my concerns are :-

• What are the performance benefits i can get?
• How real-time is it compared to Elasticsearch?
• Any additional functionality offered by Prometheus?
• Why a Timeseries Database? How does it work internally (Prometheus)?

I wish to know the cores of Prometheus and Trade-offs associated. So far what I've read online is pretty much filled with adjectives rather than some substantial evidence or intrinsic details. Some valuable resources with brief and clear details would also be great.

Thanks!

[Brian Brazil]

ELK and Prometheus are fundamentally designed for different things. ELK is event logging, while Prometheus is metrics-based. You need both types of system. This talk explains more of the very high level differences: https://www.youtube.com/watch?v=hCBGyLRJ1qo

--

Brian Brazil

www.robustperception.io

[jkim....@gmail.com]

Elastic does support metrics in addition to log aggregation.  However, in general, I would recommend using tools that do one thing and do it well.  So use Prometheus for monitoring, something like ELK for log aggregation.  That said, some enterprises do like to use a single tool, vendor etc unless there is a clear benefit.