Jaxons wrote:Can any functional test be converted to monitoring use?
In general, yes.
Any functional test can serve in a monitoring role. That's what monitoring in the eValid case really is all about: running a functional test to measure something.
However, in most cases the functional test needs some "provisioning" to assure that it is a good test to use for monitoring.
Those special steps involve things like:
(1) Adding in
ElapsedTime or
SaveRecord commands to the script to collect detailed timing data.
(2) Making sure that the
Wait times [if there are any] are of appropriate length: you want to simulate a average user, not a super-fast or super-slow one.
(3) Picking the right kind of device to tell eValid to imitate -- so your eValid settings will produce the right response from the server.
(4) Making sure that there are no problem pages in the script that might cause the playback to de-synchronize. We generally do this by attempting to force a de-sync by making the wait and delay times very small (or even zero) and then adding any needed synchronization steps. In monitoring mode you really DO want the script to be as "self synchronizing" as it can be, so that any failures detected are failues of the web site and not of the playback agent.
(5) Confirming operation from the batch file, the one that is run from the scheduler.
These investments in the script may take a bit of time, but they contribute to a very good, very reliable, very informative and useful monitoring outcom.
eValid Monitoring Team