We have a requirement to add approximately 300+ MySQL nodes to a PMM server for monitoring purposes. While reviewing the FAQ link provided below, I found some details that require clarification:
Question 1: Regarding “Memory: 2 GB per monitored database,” does this mean we need 2GB for each individual MySQL instance to be monitored? For example, for 100 MySQL nodes, would we need 200GB of memory? Is this not a significant amount?
Question 2: As mentioned earlier, could you please provide assistance in determining the total resources required (CPU, memory, disk, etc.) to monitor approximately 300+ MySQL nodes while retaining the data for a month(including QAN data as well)?
I am going to summarize the response for both questions below.
About minimum requirements documentation says:
Memory
A minimum of 2 GB per monitored database node. The increase in memory usage is not proportional to the number of nodes. For example, data from 20 nodes should be easily handled with 16 GB.
Note that there it says it is not proportional to number of nodes and hints another estimate of 16G/20 nodes. This way you would be able to monitor 100 instances with 80G.
Review following posts and responses:
Also we have an old blog but it should surely give you a good enough understanding to estimate: