kube-arangodb/docs/design/scaling.md

Scaling

The internal process followed by the ArangoDB operator when scaling up is as follows:

• Set CR state to Scaling
• Create an additional server Pod
• Wait until server is ready before continuing
• Set CR state to Ready

The internal process followed by the ArangoDB operator when scaling down a dbserver is as follows:

• Set CR state to Scaling
• Drain the dbserver (TODO fill in procedure)
• Shutdown the dbserver such that it removes itself from the agency
• Remove the dbserver Pod
• Set CR state to Ready

The internal process followed by the ArangoDB operator when scaling down a coordinator is as follows:

• Set CR state to Scaling
• Shutdown the coordinator such that it removes itself from the agency
• Remove the coordinator Pod
• Set CR state to Ready

Note: Scaling is always done 1 server at a time.