Forum Discussion

remoted's avatar
remoted
Active Contributor
7 years ago

Tips for Large Accounts

I'd like to get some input from anyone that has a large account. I have 8000 units currently and I'd like to hear any helpful information to help navigate problems with Central when using Large accounts. If there is valuable input, I'll aggregate it here.

 

  • Split groups into < 1000
    • I find that 800 is a good number for groups. It seems to be the best compromise between ease of use and responsiveness. Any more and I start to get browser failures. Smaller becomes a headache for managing the total units.
  • Use the desktop client
    • If you need to connect to a machine, but not for O2M or alerts, use the desktop client. It is much more responsive than the web site in general, especially when opening multiple connections at once. 
  • Internal Error (19)
    • With such a large deployment, this is bound to happen a lot. 

 

I'm open to any other ideas. Here are some that I'm curious about:

  • Better browser? Which browsers are you using and for what reasons (related to LMI Central)
  • Better group management
  • Management of failed O2M tasks

 

  • Adriana11's avatar
    Adriana11
    GoTo Contributor

    Hi there.

     

    You have great questions! 

     

    I would recommend getting in touch with your Account Manager at LMI and check with them if your account qualifies for a Customer Success Manager help.

    Customer Success Managers help by sharing best practices to achieve what you are looking for.

     

    Here it goes my 2 cents for O2M:

    • The default quota for O2M is 1GB/task max AND 3GB/account max. For example, users can either have 3x1GB tasks or 255x11.7MB task, up to 1000
    • The maximum number of Plans per account is 255.
    • 1k machines at a time
    • I would schedule tasks  for groups of 200/300 machines to run the plans at different times
    • You can also create a self-healing action using alerts, whenever an alert triggered the associated task will run automatically, this is great to basic troubleshooting or start services 
    • Pull the failed tasks history, and try to understand if there is a pattern to cause the same error on several machines 

     

    For reporting:

    - as you already figured out smaller groups are more responsive, I would break them into smaller groups so when running report do it by computer group, os it won't take too long to run a report

    -