We have a client to whom we need to send monthly metrics around our websites' load times based around a specific millisecond-based SLA; while we have numerous polling sites assigned to these "Web Page Speed (Browser)" monitors, following a support call with Site24x7 technical support to address erroneous data represented by our main polling site, we learned several things about this check:
1) Unlike applying maintenance windows/removing outliers for "Availability SLA", is not possible to do so for "Response Time SLA" - so in the case that a web page's speed was affected by factors during maintenance or a possible issue with the polling site, there is no way to account for this negative impact to the SLA.
2) Following our conversation, Response Time SLA is measured by the primary site polling alone, so if the primary site in question were affected, the other monitors' data is not accounted for/averaged.
Because of this, our "Response Time SLA" is not truly representative of the combined data, and we are unable to account for maintenance/factors beyond our control - which leaves us with a report that does not truly reflect our website's performance. This results in a report that does not line up with our clients' expectations.
Ideally, much like the Availability SLA (which is a reflection of maintenance windows that can be added/edited after the data was polled), we need to have some sort of control if/when data is not accurate, or is a direct result of a known issue/factor.
Thank you very much!