Highly-Available (HA) services are services deployed in configurations that allow them to better deal with unexpected stability issues. A service can go down for a number of reasons - application fault, memory issues, deadlocks, or host failures (Stratum does pretty well, but no hardware is immune to failure).
At a high level, making a service HA means to scale it up - that is, to set the number of deploy jobs. For specific service types, it gets more complex.
HA Code Services
An HA code service simply is configured to have two or more deploy jobs running simultaneously. During the provisioning of an HA code service, Catalyze will also increase the scale of the service proxy. Traffic will be routed via round-robin from the environment’s load balancer between service proxies, which will in turn only route to running deploy jobs containers for that service.
HA Database Services
HA database services vary in setup depending on the type of database.
For details on how a database type not listed above could be made HA, contact us.
HA Cache Services
HA cache services vary in setup depending on the type of database.
For details on how a cache type not listed above could be made HA, contact us.