Percentage of time that a request for a page is returned within three seconds from on campus (target: 99.9%)
* This drop was caused by a short period of downtime caused by a disk failure on a server at 4:40am, which the server then recovered from through hardware resiliency.
** This drop was caused by scheduled network maintenance from 6am to 6:10am on two mornings.
Availability off-campus, as measured by Pingdom.com (30 second timeout, from at least 2 out of 10 locations ) (no target set)
Percentage of time that a request for a websignon UserLookup request is returned within 5 seconds from within our datacenter* (target: 99.5%)
* This might seem contrived. But 5 seconds is the threshold at which other applications will time-out a request to websignon, and since the vast majority of WSOS-enabled apps live in our datacentres, that's where we should measure from.
** Downtime in October '11 caused by a resource capacity issue with unexpectedly large numbers of new students attempting to sign in, causing a 15 minute period of degraded service