Monitoring Running Container - Best Implementaion

Hi all,
I was wondering if someone can share a bolg/github repo, etc that shows the final implementation of how CheckMk is monitoring running containers in a Swarm nodes. I was able to set everything up and use PiggyBack to create a host that represents containers running in the three nodes that my Swarm cluster has. Here are the steps that I did so far:

  • Install Linux Agent and mk_docker plugin in all nodes.
  • Configure the three hosts that represents the three Swarm Nodes: Manager and workers.
  • Using Host Mapping I was able to map the piggybacks to a unique host: eg: webapp
  • The running containers are: webapp.1.{id}, webapp.2.{id}, and webapp.3.{id}
  • I create a new host called webapp
  • I was able to see that the source is comming from three sources now.

My question are:

  • Is this a good implementation.
  • If the values are coming rom three different sources, what are the final results. I feel this should be done differently.
  • I feel that I need to monitory by container name; webapp and then I need to see per host the running containers and their corresponding stats. How can I implement that?

Thank you in advance

Hallo,
first a question:
Are you using static containers with a long uptime or containers that are started for some minutes / seconds. For exampe software testing,

Ralf

It’s a constant containers running on a Swarm Cluster.

I think the problem here is that it is a Docker Swarm :wink:

If you ve the option to use kubernetes you should give it a chance.
One big advantage is that you get rid of docker on the server side without loosing the option to use it von client side.
Perhaps rancher 2.x is a way to make first steps and to evaluate what is possible using kubernetes .
Ralf

Unfortunately, Kubernetes is not an option for the moment. Let me add more details on what I am looking for. After configuring the Swarm nodes and then creating a piggyback config to map all similar containers to one single hosts, I charts I have are the followings:

My questions are:

  1. Is this how we should configure monitoring the container?
  2. The host is being fed from two different sources: RO-T-DOC02 and RO-T-DOC03. Which is Ok, but how can I make the charts show the two sources separately. Like a line for source 1 and source 2. Thee CPU utilization eg it shows one single value. Looks like both sources are mixed, correct? How can I make it showing two lines at the same chart? I am also interested in creating alerts on these values.

  1. I am also interested in knowing how services are working? Are all services up and running?

Thank you in advance

perhaps take a look here:
image

The option is not available for some reason in the 1.6 version. Do you happened to know why?

Thanks

From version 1.6.0 the CEE Checkmk Enterprise Editions supports the Dynamic Configuration Daemon – or DCD for short.

Do You use RAW or CEE Edition ???

The RAW version. I installed it from DockerHub.

then the DCD is not present in the RAW… it`s one of the Enterprise features …

What are the open point`s for you ??