We highly recommend that you setup performance monitoring (CPU, RAM, and Disk, at a minimum) on the servers that make up your Scalr cluster.
These are just guidelines. Obviously, your hardware specifications should depend on the size of the cluster you intend you deploy.
We encourage you to consider starting with something smaller, and scaling up / out as the need arises.
The following are production hardware specification guidelines for the various components that make up Scalr.
- Proxy Server: 2 CPU, 4 GB of RAM
- App Server: 2 CPU, 4 GB of RAM
- Load Statistics Server: 4 CPU, 8 GB of RAM, 100GB of Disk (SSD is preferred)
- Worker Server: 4 CPU, 8 GB of RAM
- Sessions Server: 2 GB or RAM
- Main Database Server: 4 CPU (8 preferred), 16 GB of RAM, 500 GB of Disk (SSD is preferred)
- Cost Analytics Server: 4 CPU (8 preferred), 16 GB of RAM, 500 GB of Disk (SSD is preferred)
If you are grouping multiple components (see: Deploying a Smaller Scalr Cluster), adjust the specifications accordingly. However, keep in mind that those specs leave significant headroom (so when grouping e.g. A, B, and C, you can safely get away with less than A + B + C).
Specifically, if you group Proxy + App + Worker + Sessions, you can use 4 CPU and 8 GB of RAM.
Please note: The default installation directory of /opt/scalr-server is not configurable. Be sure to mount any application storage volumes to /opt/scalr-server if they are intended for Scalr application storage.