High Availability (HA)

General definition

High availability is achieved through architectural redundancy: no single component whose failure can take the entire service offline. This applies at every layer, including application servers, databases, message brokers, load balancers, and network paths. When one instance fails, traffic shifts automatically to healthy instances, and users experience only a brief latency spike rather than an outage.

• Active-active: all nodes serve traffic simultaneously; any node can fail without service interruption
• Active-passive: a primary node serves traffic and a standby node takes over on failure
• Health checks and load balancers detect failures and route around them automatically
• Database replication ensures no data loss when a primary database node fails

HA architecture directly enables the high uptime percentages stated in SLAs and SLOs. It is distinct from disaster recovery, which addresses recovery after catastrophic or regional failures rather than the routine single-node failures that HA handles automatically.

In the Ethora ecosystem

On a cloud-hosted Ethora plan, high availability is built into the shared infrastructure. The provider operates redundant nodes, manages failover, and publishes the resulting uptime commitment in the SLA. Customers benefit from HA without designing or operating the architecture themselves.

With a dedicated or self-hosted chat server, you own the HA architecture. Ethora’s platform is built on ejabberd, which supports clustered deployments across multiple nodes and regions. The Ethora team can design a multi-node cluster for your environment, configure health checks and load balancers, and set up database replication to match your SLO target. For healthcare and financial-services deployments this means demonstrating to auditors exactly how availability is achieved and tested, an advantage over a black-box managed service. Contact the team to scope a dedicated HA deployment.