Aggregate prediction outcomes

So far, we've looked at a lot of insights about individual servers. But what if we need to look at groups of servers as a whole?

For example, you might have a rack of 24 servers, each with a 10 Gbps connection and a top-of-rack network switch that can aggregate a total of 40 Gbps of bandwidth. That means that your total rack network activity could be 240 Gbps, but you must squeeze it all down a 40 Gbps connection. Wouldn't it be good to know when your rack is predicted to approach its 40 Gbps limit?

Alternatively, you may have a cluster of storage servers. You're not too concerned with each individual server's disk usage, but you are interested in the cluster as a whole. You want to know when the total space on all your servers is predicted to pass through a threshold so that you can add more storage or purge old files.

We can use the lessons we've learned in the previous recipes (and previous chapters!) to start looking...