DX Unified Infrastructure Management

  • 1.  How to Add a Robot in multiple groups to have different Disk monitoring profiles

    Posted Nov 28, 2018 09:18 AM

    Hi All,

     

    We have a requirement where a server containing the multiple mount points and need to have different thresholds for different mount points.

     

    Example:

    Below mount points should be in Global Threshold Group with Aggressive Thresholds: >= 70% Major and >=80% Critical

    /app

    /opt

    /move

     

    Below mount points should be in Exception Group with Normal Thresholds: >=90% Major and >=95% Critical

    /tmp

    /

    /user

     

    I don't want Spend time on configuring the each device and each mount point on almost 800+ devices. Please let me know how can I apply the above kind of multiple thresholds to a group of devices using MCS templates or any other options. 

     

    Thank you.

    Rajashekar

     

    Gene_Howard



  • 2.  Re: How to Add a Robot in multiple groups to have different Disk monitoring profiles

    Posted Nov 28, 2018 03:52 PM

    Using infrastructure manager you could create a CDM archive package with those settings predefined. Then deploy that using the traditional method.

     

    Gets trickier if the filesystems don't have completely predictable names though your query seems to indicate all 800 devices have the same filesystems.



  • 3.  Re: How to Add a Robot in multiple groups to have different Disk monitoring profiles

    Posted Nov 29, 2018 05:13 AM

    Thank you Garin. Here the problem is most of the servers do not have the same/identical names for file systems and I would like to use dynamic thresholds instead of static thresholds.

     

    Any other way that we can use to achieve the above requirement?



  • 4.  Re: How to Add a Robot in multiple groups to have different Disk monitoring profiles

    Posted Nov 29, 2018 11:27 AM

    Maybe you could add the mental process you would go through to establish the thresholds manually. UIM is not going to provide a decision process or method for you that's going to automatically do what you want unless you start off knowing exactly what you want. In this case it appears that you are requesting help to eliminate the laboriousness of a manual configuration process - so it's really helpful to start off that question with clear detail on that manual process you are trying to replace.

     

    At a minimum it is possible to code in Lua (and a bunch of other supported languages)  pretty much anything that you could do manually. The problem is that Lua doesn't know what you want unless you can supply the decision logic. Look into the controller callbacks - get/set config will get you started.

     

    The archive package manager is extremely powerful with regards to managing and deploying configurations - it's one of the hidden gems in the product that CA doesn't seem to know or understand they own. Something that might help you here is that fact that when you deploy a traditional package, there's opportunity to supply a command that runs before the install section and one that runs after. So you can create a package that includes a default CDM cfx file and a batch file that figures out what drives are there and what their configuration settings should be and then, in the post install command section, it can run the batch file to apply those.

     

    The MCS tool is on attempt three or four or five, depends on how you count the beta/alpha/PoC, etc releases along the way, to provide a template based configuration tool. So far all the attempts in my opinion have fallen well short of the existing archive package toolset because there's no true flexibility. You can coerce it to some extent but then what you are doing there is putting all the manual configuration labor into the templating process on the central server and pushing that out instead of applying a similar amount of labor to the remote system. It does work for some situations but not many. It's great if you want to get CPU monitoring on everything so you can put up graphs.



  • 5.  Re: How to Add a Robot in multiple groups to have different Disk monitoring profiles

    Broadcom Employee
    Posted Nov 29, 2018 12:27 PM

    The Disks MCS template creates a separate profile for each file system which means you set the two thresholds per file system.

    So you would have 3 profiles for /app; /opt; /move which have one set of thresholds and the others, /tmp; / ; /user will have different thresholds.

    Create these profiles at the Group level and control the rollout of these profiles by group membership (dynamic on the content of a user_tag or static for manual selection)

    Would this work or have I missed something?

    Cheers

    Rowan