Instana

Instana

The community for performance and observability professionals to learn, to share ideas, and to connect with others.

 View Only

  • 1.  Best Practices: Monitoring the Instana backend

    Posted 08/26/25 09:22 AM

    Hi,

    Instana is powerful and fast but needs resources to enable it. We (bigger IT organisation in Germany) are currently on the mission to setup a big custom self-hosted installtion on OpenShift and we need a little guidance.

    One of the things we need is a way to monitor the backend components so we can take appropriate measures before something serious happens. We lack the experience though and are interested in existing best practices, dashboards, etc.

    Even hints and tips on what to look out for are greatly appreciated. 

    I am sure that we are not the only ones out there with questions around operating a huge Instana installation and hope for a vital interest and lots of participation here.

    Thanks and regards,

    Rainer


    #Self-Hosted

    ------------------------------
    Rainer Schuppe Rainer Schuppe
    Instana veteran (since 2016), APM oldtimer (since 2003 Wily, Dynatrace, AppDynamics, Instana, OpenTelemetry)
    https://observability-heroes.com
    ------------------------------


  • 2.  RE: Best Practices: Monitoring the Instana backend

    Posted 09/05/25 03:30 AM

    Bump ... No one?

    I am a little disappointed to be honest. Not even a hint from the admins or moderators - There's got to be somebody else out there who runs Instana in a bigger environment.

    Sharing is caring :-)

    Regs,

    Rainer



    ------------------------------
    Rainer Schuppe Rainer Schuppe
    Instana veteran (since 2016), APM oldtimer (since 2003 Wily, Dynatrace, AppDynamics, Instana, OpenTelemetry)
    https://observability-heroes.com
    ------------------------------



  • 3.  RE: Best Practices: Monitoring the Instana backend

    Posted 09/05/25 09:39 AM

    Within the SWAT/Tiger team we are doing some work to create some custom dashboards and Events/Smart Alerts.   But, it is a work in progress.  Once we have a reasonable amount complete, we will publish.  

    Step 1 would be to setup the out of the box sensors for all of the key technologies that are used by Instana.  Setup the sensor for cassandra, postgres, clickhouse, etc.

    There are some out of the box Events for most of these technologies.

    Step 2:  Make sure you setup a Smart Alert to monitor the PVC space available.  Keep in mind that Elasticsearch starts to have problems when it reaches around 85% capacity in the PVC, so set your threshold significantly lower than 85%.   The other datastores will utilize the full capacity of the PVC.



    ------------------------------
    Ben Stern
    Executive IT Specialist / Monitoring and Analytics Best Practices and Technical Evangelist
    IBM
    Durham NC
    919-224-1593
    ------------------------------

    Original Message


  • 4.  RE: Best Practices: Monitoring the Instana backend

    Posted 09/08/25 06:30 AM

    Hi Rainer, we have the same problem.

    But we are not the only ones.

    There is an idea in the portal 

    Self-Monitoring Instana with Instana - Document Instana custom metrics [INSTANA-I-1684]

    that was created in July 2023 and there is no solution til now. Please vote it. It is not a my idea but I believe that it so important.

    I think that Instana SWAT/Tiger team should do better ...

    Carla 



    ------------------------------
    Carla Raffi
    System programmer
    Sogei
    +393387889337
    ------------------------------



  • 5.  RE: Best Practices: Monitoring the Instana backend

    Posted 09/08/25 06:49 AM

    Hi Carla,

    Thanks for the link, I voted for it.

    I don't consider it rocket science to provide at least some guardrails for such a complex system as Instana. I love the product but it needs to be manageable.

    Regards,

    Rainer



    ------------------------------
    Rainer Schuppe Rainer Schuppe
    Instana veteran (since 2016), APM oldtimer (since 2003 Wily, Dynatrace, AppDynamics, Instana, OpenTelemetry)
    https://observability-heroes.com
    ------------------------------

    Original Message


  • 6.  RE: Best Practices: Monitoring the Instana backend

    Posted 09/09/25 10:27 AM

    Hi Rainer, we have a similar requirement and an associated IDEA as well (IBM Ideas).

    We spoke with a group of Instana developers and they recommended to have at least 2 Instana instances so that they could monitor each other.  From what I gather this is what they do with their SAS implementation of Instana.  From my understanding if Instana is monitoring itself it creates a type of feedback loop which will eventually consume all of the available resources.



    ------------------------------
    Dale Falardeau
    ------------------------------



  • 7.  RE: Best Practices: Monitoring the Instana backend

    Posted 09/10/25 03:35 AM

    Hi Dala, thanks for that info. 

    I know that it is vitally important to run the agent which is monitoring Instana only in INFRASTRUCTURE mode, as the APM mode will overwhelm the backend with useless traces. 

    And that's exactly why we are looking for a best practices advice, as we are going to run a huge setup and need to be able to manage it here.

    Regs,

    Rainer



    ------------------------------
    Rainer Schuppe Rainer Schuppe
    Instana veteran (since 2016), APM oldtimer (since 2003 Wily, Dynatrace, AppDynamics, Instana, OpenTelemetry)
    https://observability-heroes.com
    ------------------------------

    Original Message


  • 8.  RE: Best Practices: Monitoring the Instana backend

    Posted 09/10/25 09:59 AM

    Yes, we've stood up separate nodes for the Instana back-end so that we can run the agent in APM mode in the rest of the cluster.



    ------------------------------
    Dale Falardeau
    ------------------------------

    Original Message