Start small and upgrade only when you need to. The Medium plan is our most popular and that is what we recommend for first-time deployments. And that’s the rule-of-thumb. If you’re looking for a more technical and more accurate approach, read on! Joe has you covered…
Varnish Cache is largely thread-driven. It spawns a number of threads to handle workload sessions. Each thread gets to handle exactly one session. The more threads that are active and available, the more sessions it can handle! And the more memory you have, the more threads you can start up. This doesn’t scale infinitely but then you don’t need several thousand threads in practice for most workloads… Maybe you do! We have you covered for that anyway; so no worries!
So let’s see, you need to handle 1 million unique visitors a month with a maximum of 200 simultaneous visitors. Let’s say that each visitor generates 5 sessions on average. For that you need a Cachoid that can handle 1000 sessions at once. In Varnish nomenclature, you need roughly 1000 worker-threads. And, roughly again, you’ll need 2GB for Varnish or the XXL plan.
For larger website that non only need to serve more traffic but also require redundancy and reliability, Smart Load Balancing will fit the bill. Picking more than one location for your cachoid will instruct Cachoid to deploy load balancing across your locations. This feature automatically distributes traffic to your cachoids work in concert to serve traffic.