Agent Health solution in OMS


Summary: The Agent Health solution is now in public preview.

Hi everyone, Nini here, and I'm following up from my last blog post. I am happy to announce that today you are now able to use the Agent Health solution (in public preview) and more easily view visualization of your agent heartbeat events and agent distribution statistics across your environment. This solution was built by using the View Designer with the queries mentioned in my Agent Health MP post. However, by enabling this solution now, you automatically get the next version of the solution with a significant number of changes.

First, search for the solution in the Solutions Gallery, and enable it:

Find the solution in the Solutions Gallery, and enable it.

After the solution is enabled, the Agent Health solution tile is added to your OMS Overview page and shows the total count of agents and the number of unresponsive agents in the last 24 hours.

Screenshot of the total count of agents and count of unresponsive agents in last 24 hours.

Select the Agent Health tile to open the Agent Health dashboard. The dashboard includes the columns in the following table. Each column lists the top ten events by count that match that column's criteria for the specified time range. You can run a log search that provides the entire list by selecting See all at the right bottom of each column,or by clicking the column header.

Agent Health Dashboard blade Description
Agent count over time A trend of your agent count over a period of seven days for both Linux and Windows agents.
Count of unresponsive agents A list of agents that haven’t sent a heartbeat in the past 24 hours.
Distribution by OS Type A partition of how many Windows and Linux agents you have in your environment.
Distribution by Agent Version A partition of the different agent versions installed in your environment and a count of each one.
Distribution by Agent Category A partition of the different categories of agents that are sending up heartbeat events: direct agents, SCOM agents, or the SCOM Management Server.
Distribution by Management Group A partition of the different SCOM Management groups in your environment.
Geo-location of Agents A partition of the different countries where you have agents and a total count of the number of agents that have been installed in each country.
Count of Gateways Installed The number of servers that have the OMS Gateway installed, and a list of these servers.

 

Agent Health dashboard

Agent Health dashboard

The following table shows data collection methods and other details about how data is collected for Agent Health:

Platform Direct Agent SCOM Agent Azure Storage SCOM required? SCOM agent data sent via management group or directly to OMS Collection frequency
Windows Yes Yes No No Yes 1 minute
Linux Yes No No No -- 5 minutes

What's next?

We are currently working on an extended version of this Agent Health OMS solution, which will provide a more in-depth visualization in OMS and suggestions to help guide you in taking actionable remediation steps.

That’s all I have for you today. Feel free to try this out in a live environment! If you have any feedback or questions, I’d love to hear from you. Please comment below.

Priscilla Nini Ikhena
Microsoft Operations Management Team

Comments (10)
  1. Dirk Dulfer says:

    Cool and useful new addition. Question about the Agent Geo-Location though.
    Is the location listed there the geo-location of, in my case, the Azure VM? If so, than I wonder why this differs from the location reported in the VMs blade in the Azure portal.

    1. Nini Ikhena says:

      What locations are you getting for both?

      1. Dirk Dulfer says:

        Only just noticed your reply.
        In OMS, I see an agent (VM Classic) supposedly located in Ireland, whereas the VM’s location is set to West Europe.
        Resource Group location is set to West Europe as well.

        All other VMs in the RG show their correct location (Netherlands), but one particular VM shows Ireland.

  2. M.mathew says:

    Am noticing “ComputerIP” values not consistent with the Agent IP’s. Any thoughts?

    1. Nini Ikhena says:

      Can you shoot me an email of a screenshot of the Agent IP address you see?

  3. M.mathew says:

    All Agents in OMS(agent Health) show up with IP of SCOM Management Server

    1. Sambit Mishra says:

      Each OMS agent reporting to management servers will be sending up two heartbeats with values of SCAgentChannel as Direct and SCManagementServer depending on whether its going directly to OMS or through the management server respectively. For heartbeat events which have SCAgentChannel as SCManagementServer the IP is set to the management server since the data is actually uploaded by it. For the heartbeats where SCAgentChannel is set to Direct you should be seeing the public IP of the agent.

  4. Nini Ikhena says:

    This is correct.

  5. Łukasz says:

    Hi,

    I would like to turn on alert when my internal page in iis in my own server fall down. How can I configure oms ?

    I configured alert in my server like this
    Type=Heartbeat | measure max(TimeGenerated) as LastCall by Computer | where LastCall < NOW-5MINUTE

  6. Rahul Shinde says:

    Hi,
    I’ve Agent Health solution add to portal, the unresponsive agent for last 24 hrs had reflected but after few days that unresponsive agent goes away. Is this normal practice that unresponsive agent doesn’t reflect? Is there any solution to this or any setting required?

Comments are closed.

Skip to main content