Services collects awesome data, but with complex systems (monitoring a few hundred devices), it can be a little overwhelming and not particularly useful for executive summary style presentations. A few feature enhancements would significantly increase the value of this data:

• Most important would be export to CSV/tab delimited data for export and further analysis. ZBXNEXT-553 would be very helpful with this feature. I'd amend that request with the addition of "all" (just like in graphs)

• Sorting by any header item would be great - sort by problem time (best to worst/worst to best) and by SLA.

• View filter: like the latest data filter (host groups, hosts, application, name) would be perfect plus check box for "only unacceptable SLA" (at the very least, limit the view to only "unaccetable SLA" services.)

• Alternate status calculation algorithm for children: average of children's problem time. This is useful for comparing one group of devices to another group.

• It appears that "no data" (or before data) is calculated as "not problem" in the calculation of SLA. For example a device that is added in a problem state has an SLA that is calculated as "time monitoring/Period." for example 24 hours after adding, the "last 30 days" will show SLA of 1/30 or 96.6666/SLA even though it has been in problem state 100% of the time it has been monitored. I'm not sure if there's a rule for this, but I would think the SLA should be 0/Acceptable SLA). It would seem that there should be at least an option to treat "no data" as "in problem state" rather than "no data is good data."