Dynatrace Managed high availability protects against different scopes of failure. A single Managed Cluster tolerates the loss of individual nodes. Premium High Availability (PHA) adds a second data center, so monitoring continues even if an entire data center fails. On top of either topology, rack-aware deployment groups nodes into fault domains, so a Managed Cluster also survives the outage of an entire rack. First choose a base topology, either a single Managed Cluster or PHA, then optionally layer on rack-aware deployment. The following pages help you compare these options, add a data center, convert an existing Managed Cluster, and recover after an outage.
Understand how Premium High Availability (PHA) uses redundancy, hardware placement, capacity planning, data replication, and automatic failover across data centers.
Replicate Managed Cluster nodes across two data centers and prepare Cassandra, Elasticsearch, and server components for PHA.
Review how PHA detects Elasticsearch and Cassandra node outages, routes failover decisions, and restores service after recovery.
Recover a data center in a PHA deployment by using another data center as the recovery source.
Recover a lost data center from backup and restore the data center topology in a multi-data center deployment.
Rebuild a lost data center, reinstall nodes, migrate data stores, and restore replication across both data centers.
Learn how rack-aware deployments group nodes into fault domains to survive a rack outage, on a single Managed Cluster or a PHA deployment.
Convert a Managed Cluster to a rack-aware deployment using the expansion method when you can add capacity before migration.
Convert a Managed Cluster to a rack-aware deployment using the restore method when you need to rebuild from restored data.