There are two options for configuring operational alerting in a multidatacenter environment:
- Option 1: Run the metrics and SingleStore alerting servers on both Site A and Site B. Disable all alerts on the Site B alert server except for Cassandra based alerts.
- Option 2: Run the metrics and alerting server on Site A, and shut down the alerting server on Site B. Ensure that the Cassandra and SingleStore node count check alert on Site A is updated to include the number of Cassandra and SingleStore nodes in Site B. This would allow for alerting on the case where the Site B Cassandra or SingleStore data center node(s) are not part of the Cassandra / SingleStore clusters for any reason, but has the limitation that other Cassandra / SingleStore metrics such as CPU, load, disk space, and compaction for Site B Cassandra or SingleStore nodes would not be monitored.