Show "Latest Outages" with Check Plugin Output in SLA Reports

[sva-ew]

Is your feature request related to a problem? Please describe.

The current SLA reports (Host SLA, Service SLA) in Icinga DB Web show only the SLA percentage per host/service for the selected time range. When a service has an SLA below 100% (or below the threshold), there is no way to see why it was unavailable — the check plugin output (e.g. HTTP CRITICAL: HTTP/1.1 502 Bad Gateway - 295 bytes in 0.012 second response time) is not shown.

Users migrating from Thruk/Icinga 1 reporting are used to a "Latest Outages" block that lists, for each outage:

• Start time – End time (and duration)
• The check plugin output at the time the outage started

This context is essential for operations teams to understand SLA violations and to communicate them to stakeholders in monthly/quarterly reports. A bare "99.96%" without the reason is of limited value when investigating incidents.

Describe the solution you'd like

Add an opt-in "Latest Outages" block to the SLA report, appended below the existing SLA table, that lists the most recent outages within the selected time range with:

Host	Service	Start	End	Duration	State	Output
host1	HTTP vHost [...]	2026-07-23 11:09:56	2026-07-23 11:27:49	0d 0h 17m 53s	CRITICAL	HTTP CRITICAL: HTTP/1.1 500 Internal Server Error - 237 bytes in 0.125 second response time

Proposed configuration options

• "Show Latest Outages" (checkbox, default off) — opt-in to append the outages block to the report
• "Max. Outages to Show" (number, default 20, max 100) — limit the number of outages shown (most recent first)

Outage definition

An outage is a hard state transition from OK (0) to any non-OK state (WARNING/CRITICAL/UNKNOWN). The outage ends at the next transition back to OK, or at the end of the selected time range if the outage is still ongoing. Transitions between non-OK states (e.g. WARNING → CRITICAL) are part of the same outage and do not start a new one; the original output is preserved.

Data source

The data is already available in the IcingaDB state_history table (output, long_output, hard_state, previous_hard_state, event_time columns). No schema change is required.

Describe alternatives you've considered

• Manual investigation: Opening the history view for each service with SLA < 100% and scrolling through state changes. This is tedious and not feasible for monthly reports covering many services.
• Custom external reporting: Building a separate tool that queries state_history directly. This duplicates effort and doesn't integrate with the Icinga Web 2 reporting framework (scheduling, PDF export, templates).
• Thruk alongside Icinga DB Web: Running Thruk just for this report. This adds operational overhead and a second monitoring UI.

Additional context

• The IcingaDB state_history table has an index on (environment_id, event_time) but not on (host_id, service_id, event_time). For large environments with long time ranges, the outage query should be filtered to only fetch outage-relevant transitions (hard_state > 0 AND previous_hard_state = 0 OR hard_state = 0 AND previous_hard_state > 0) rather than all hard state changes, to keep the result set manageable.
• History retention may limit the availability of older outage data. For "Last Year" reports, outages older than the retention period will not be shown. This is a data constraint, not a code issue — the same applies to the SLA calculation itself.
• The related (but different) icingaweb2-module-idoreports PR Fix incompatible hook usage with 3rd party modules #45 "Add report Host Outage" (open since Feb 2021) adds a cumulative outage-time column for the old IDO-based reporting, but does not show individual outage events or check output.

Mockup

┌─────────────────────────────────────────────────────────────────────────────┐
│ SLA Report: Service SLA                                                     │
│ ┌──────────────┬──────────────────────────┬─────────┬────────┐              │
│ │ Hostname     │ Service Name             │ Month   │ SLA in %│             │
│ ├──────────────┼──────────────────────────┼─────────┼────────┤              │
│ │ host1        │ HTTP vHost [checkbuy...] │ 2026-07 │ 99.96  │              │
│ │ ...          │ ...                      │ ...     │ ...    │              │
│ │ Total (180 Services)                    │         │ 99.87  │              │
│ └──────────────┴──────────────────────────┴─────────┴────────┘              │
│                                                                             │
│ Latest Outages                                                              │
│ ┌────────┬───────────┬─────────────────────┬─────────────────────┬─────────┬──────────┬──────────────────────────────────┐
│ │ Host   │ Service   │ Start               │ End                 │ Duration│ State    │ Output                           │
│ ├────────┼───────────┼─────────────────────┼─────────────────────┼─────────┼──────────┼──────────────────────────────────┤
│ │ host1  │ HTTP vHost│ 2026-07-23 11:09:56 │ 2026-07-23 11:27:49 │ 17m 53s │ CRITICAL │ HTTP CRITICAL: HTTP/1.1 500...   │
│ │ host2  │ ping4     │ 2026-07-15 08:02:11 │ 2026-07-15 08:05:44 │  3m 33s │ CRITICAL │ PING CRITICAL - Packet loss = ...│
│ └────────┴───────────┴─────────────────────┴─────────────────────┴─────────┴──────────┴──────────────────────────────────┘
└─────────────────────────────────────────────────────────────────────────────┘