Most people argue that basic server metrics to monitor is CPU load, free memory & swap, disk space, disk I/O, and network/bandwidth I/O. Even if Topbeat include & extend some of these metrics it is also missing some.
So Is the roadmap for topbeat aiming to extend collected metrics or should on one focus on develop beats addressing missing "features"?
Currently the roadmap is not fixed yet. The plan for topbeat is to keep it small as it should also be an example on how to create a simple beat.
For the metrics we are thinking about creating something like a metricbeat or sysbeat to gather more metrics. As we will never be able to gather all possible metrics, we are thinking also of an extendable model.
The advantage of having lots of small beats is that everyone can pick the features he needs. The disadvantage is that probably from a certain number of beats it gets hard to manage all the beats.
I would currently recommend to create multiple small beats for the missing features. This will also allow us and the community to iterate and learn fast on small projects and we will see what works and what doesn't. We already see it now with topbeat, that getting a first example running is very easy and can be done in a few hours, but then make it compatible with all platforms and finding the edge cases is hard.
We already now have the discussion if windows event log beat should be part of filebeat or a separate beat. As said in the beginning, it is not decided yet.
In case you create your own beat, make sure to add it to the list here: https://www.elastic.co/guide/en/beats/libbeat/current/community-beats.html Of course also contributions to topbeat are more then welcome.