There is a need to consider SLA on the zabbix side. But at the moment, all the problems in calculating the SLA have the same weight. The main service or subordinate agent has stopped - the SLA will be calculated for them in the same way, but in practice, when the main service is stopped, we have complete inaccessibility or significant degradation of the service, which should have a greater impact on the SLA.
Suppose we want to consider infrastructure SLA services by means of zabbix, we have 100 agents, a zabbix-server and a database. We put all this in the Services and we get that if 1 agent is unavailable, our SLA decreases just as if the database is unavailable. BUT, if the database is unavailable, we believe that the entire service has degraded by N%.
I would very much like to be able to set coefficients when configuring SLA