Werk #22689: BI availability: n/a instead of actual state for unreachable 'Count OK' levels
| Component | BI | ||||||
| Title | BI availability: n/a instead of actual state for unreachable 'Count OK' levels | ||||||
| Date | Oct 16, 2026 | ||||||
| Level | Prominent Change | ||||||
| Class | Bug Fix | ||||||
| Compatibility | Compatible - no manual interaction needed | ||||||
| Checkmk versions & editions |
|
If a BI rule uses the aggregation function Count the number of nodes in state OK and its levels require more OK nodes than the rule has child nodes, the aggregate can never be OK. Checkmk correctly computed CRIT (or WARN, if only the OK level was out of reach) for it, but also considered the aggregate to be out of its service period. As a result, the availability timeline showed these phases as n/a instead of CRIT or WARN. Depending on the levels, the aggregate was also considered to be in a scheduled downtime.
This mostly affected the history: older phases often have fewer child nodes, for example because services were added to the aggregation later.
Such aggregates are now reported with their actual state. They are only considered out of their service period or in a downtime if their child nodes are. If both levels are out of reach, so that the aggregate is always CRIT, this requires all child nodes to be out of their service period or in a downtime.
The same applies to the live BI views, which no longer wrongly show these
aggregates as in downtime or out of their service period. If you use the
special agent rule BI aggregations with the option State, if BI
aggregate is in scheduled downtime, the affected Aggr services were
reported with that state instead of their actual one. After the update,
they show their actual state and may turn CRIT. You don't need to change
anything.