Simplify minimum resource requirements for Big Bang
Currently, the Big Bang Excel spreadsheet is used to keep track of minimum required memory and CPU based on the applications deployed. We need to consider the requested resources, replicas, scaling, load and feature set. For feature set, we can cover the "typical" sets only (e.g. default, enterprise). I think the following table should cover the parameters we need. Note that prerequisites, like flux, should be included. But, Kubernetes distros should not.
Memory (in Mi)
Package | Configuration | Fixed | Variable | Replica Method | Notes |
---|---|---|---|---|---|
SampleApp | Default | 50 / 128 | 100 / 256 | Fixed (3) |
Configuration = the feature set that is turned on/off (e.g. enterprise vs. free)
Fixed = the amount required in Big Bang based on requests for resources w/o replicas. The first number is the base required (e.g. Big Bang requests), the second is the heavily loaded value (e.g. Party Bus).
Variable = the amount required in Big Bang based on requests for resources w/ replicas. The first number is the base required (e.g. Big Bang requests), the second is the heavily loaded value (e.g. Party Bus).
Replica Method = One of the following: Fixed (default), Scaled (min,default,max), Per Node
There may be situations where multiple variable methods exist. Notes can be used to explain this.
Similar tables would be used for CPU and persistent disk space
The table above is a suggestion and can be modified to make things more intuitive.