Track mesos run time stats with Metricbeat

Hello - I have configured metric beat to read container info from Docker containers. I want to now capture mesos cluster runtime metrics into metric beat. Is there a way to do this? I used the docker.sock file for metrics.

Can you tell me more about Mesos? How does Mesos expose this information? Is there an HTTP endpoint for the node that exposes this data? Probably we could add a mesos module to capture this data.

What types metrics would it collect? Have examples?

Here is the info on Apache Mesos Metrics:

Do let me know if this helps. Thanks for responding.

It's great that the API is documented so well. It would probably be pretty easy to implement a new module to monitor those metrics. Please open a new issue in the Github repo to track this enhancement request.

If you are interesting in contributing we have a dev guide for Metricbeat. You could use the golang/expvar metricset as an example because it does nearly the same thing -- fetch JSON from a REST endpoint.

+1 on a mesos metricbeat module. In case you don't have time for a PR yourself, perhaps you can open an enhancement request on github with all the data / links above.

Is it possible to extract that info via Zookeeper? I see a module of Zookeeper in Metricbeat and Zookeeper takes care of scheduling container within Mesos/Marathon.

The zookeeper modules it to monitor zookeeper itself, not to retrieve data. But we have a similar case for the Kafka module where some data must be (indirectly) fetched from Zookeeper. I'm ok if a kubernetes module would connect to zookeeper to get the meta data in case zookeeper is always around inside kubernetes.

@ab-agni, Have you created an issue to track this ? I'd be willing to help to get this implemented since we need a similar module for tracking our mesos stats.

This topic was automatically closed after 21 days. New replies are no longer allowed.

@surajs There is no Github for a metricbeat mesos module yet. Feel free to open one on Github.