• State Not Answered
  • Replies 2 replies
  • Subscribers 59 subscribers
  • Views 292 views
  • Users 0 members are here
Options
Related

Foglight for Databases - Federation Parent Hardware Requirements

dawid mandyna
22 days ago

Hey,

I'm looking to deploy a Foglight with Federation enabled, but I'm struggling to find hardware requirements for the Federation master Management Server.

Let's say I've got 5 federated children with 100DBs each, does the Federation Master need to meet the requirements to support 500DBs, or is most of the processing covered by each of the Federated Child, and the Federation Master requires less processing power?

I would really appreciate if you could either point me to the documentation or let me know how we can calculate the hardware requirements for the Federation Master.

Kind regards,
Dawid

Parents
  • 0 22 days ago

    Hi Dawid,

    I looked in the Federation Field Guide on support and it refers to the System Requirements guide for the federation master. The "single tier" requirements show 4 cpu and 12 GB memory. The federation master is really a query node - it sends requests to the child FMS' for their data, etc. The bulk of the processing is done on the child nodes. The other factor to consider is network latency between the parent and child nodes. I would make sure the individual FMS' doing the monitoring are sized properly for 100 instances, then start with a 4 core / 16 GB parent. Measure the response time in the UI.. adjust from there.

    Regards,

    -Darren

Reply
  • 0 22 days ago

    Hi Dawid,

    I looked in the Federation Field Guide on support and it refers to the System Requirements guide for the federation master. The "single tier" requirements show 4 cpu and 12 GB memory. The federation master is really a query node - it sends requests to the child FMS' for their data, etc. The bulk of the processing is done on the child nodes. The other factor to consider is network latency between the parent and child nodes. I would make sure the individual FMS' doing the monitoring are sized properly for 100 instances, then start with a 4 core / 16 GB parent. Measure the response time in the UI.. adjust from there.

    Regards,

    -Darren

Children