How frequently are health checks done on the Webpulse service?
How frequently are health checks done on the Dynamic Real-Time Rating Technology (DRTR) included in the SG appliance?
Note: Health checks include DNS lookup of the DNS name sp.cwfservice.net for SGOS 6.5 and webpulse.es.bluecoat.com for SGOS 6.6. and connectivity checks.
Unlike the K9® or ProxyClient, the ProxySG stays with a certain IP address until its health has been discovered suspicious. The client (K9, proxy client) does a health check every 100 Category checks.
By default, the ProxySG checks the health every 10 seconds. In a healthy state, the ProxySG checks every 10800 seconds. Both are user-configurable.
The ProxySG will only change the selected service point after a failed health check. N failures will cause an early health check to 'trigger'. By default, N = 1. This is configurable. (That is, if we are in our long 10800-second healthy window, a WebPulse failure will trigger an immediate health check. That may cause us to select a new service point.)
DNS resolution is managed by Health Checks, using its background DNS refresh settings. These use the TTL from the DNS response but is subject it to a user-configurable ceiling and floor. By default, the floor for the TTL is 30 seconds, and there is no hard TTL ceiling. There is a DNS negative result TTL of 10 seconds (configurable). The TTL for a record resolves for sp.cwfservice.net or webpulse.es.bluecoat.com are ~30 seconds.