OP5 SLA calculation method: Cluster mode (best state)

There were instances where OP5 users are creating Availability Reports and generate an SLA for any redundant services (for example: redundant switch devices) setting up “SLA calculation method (best state)”.

As the two redundant services doesn’t have downtimes on the same timeframe, the Admin is expecting a 100% SLA. However, the result of the report is otherwise, yielding something like 99.995% instead.

This is because “Availability vs. Downtime defines that the entire cluster is only counted as "Down" (unavailable) during periods when all monitored nodes are in a non-UP state (e.g., CRITICAL, WARNING, UNKNOWN, or DOWN). For example, the PING service triggered “WARNING” states due to packet loss. Since WARNING counts as non-UP, overlapping WARNING periods on multiple hosts particularly on redundant service/host (whether a network device or VM instance) may affect the cluster’s calculated availability”.

Please also note that the actual SLA value may vary depends on your “State type” settings either “Hard” or “Soft”, or both.