By default, we attempt to connect to your site from three different monitoring agents using the three-strike rule. Finally, before we send out an alert, we do a sanity check against a known good site to verify the agents are working as expected.
By confirming that the error is occurring in more than one location, and double-checking that our agents are able to connect to other sites, this helps rule out potential agent issues. In diagnostic alert emails for connectivity or timeout errors, traceroute information is also included to help understand and/or diagnose issues that lie along the route between our agent and your site.
We continually monitor the health of our network and review performance information to ensure agents are working as expected. We will pro-actively remove an agent from the network if there appears to be a potential issue, so it can be diagnosed or upgraded as necessary.