You can determine a connection failure based on whether a backend target is not responding, or on specified HTTP status codes that should be treated as failures of a specific backend target. For example, if a backend target is responding to requests with a 500 Server Error status, you might consider that server down even though the web service is responsive.

If you specify multiple targets in a site configuration but a load balancing strategy is not applied, the availability profile will cause the first listed target in the site configuration to be used unless it fails. Secondary targets are only used if the first target is not available.

When you change the availability profile for a site, the new availability profile configuration only applies to new connections. Existing connections continue to use the previous availability profile configuration. To apply the new availability profile configuration to all connections for a site, you must restart the PingAccess engines.

Currently, the only availability profile type is On-Demand. You might want to create different profiles for different sites based on your site needs for retry counts, retry delays, timeouts, or HTTP status codes.