Skip to main content
Version: Next

SLA API

SLA endpoints report task deadlines, escalation events, and administrative metrics. Task responses include nullable dueDate, slaStatus, slaTimeZone, and completedWithinSla. Tasks with no deadline have no SLA status. Status values are ON_TRACK, AT_RISK, OVERDUE, and COMPLETED.

Operations​

MethodPathSummary
GET/admin/sla/metricsgetSlaMetrics

Summarize visible tasks​

POST /tasks/sla/summary accepts the same TaskSearchRequestDto filters as task search. Requires ACCESS_BPM_ADMIN, ACCESS_PROCESS_PORTAL, or SCOPE_tasks:read. Administrators see all matching tasks; other callers are limited to visible tasks.

curl -X POST "http://localhost:8080/tasks/sla/summary" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"filters":[]}'

Example 200 OK response:

{"onTrack":3,"atRisk":1,"overdue":2,"completed":4,"withoutSla":5,"asOf":"2026-10-16T12:00:00Z","timeZone":"UTC"}

Task search also accepts dueDate comparisons with offset-qualified timestamps and slaStatus filters. Use the configured SLA time zone when presenting deadlines.

Read task SLA events​

GET /tasks/{id}/sla/events?page=0&size=20 requires task read access and checks task visibility. An unavailable or invisible task returns 404. It returns a page of SlaEventDto, ordered by occurrence time descending, with event outcome, scheduled and actual times, deadline, and optional incident ID. Page size is capped at 100.

curl "http://localhost:8080/tasks/123/sla/events?page=0&size=20" \
  -H "Authorization: Bearer $TOKEN"

Administrative metrics​

GET /admin/sla/metrics requires ACCESS_BPM_ADMIN. groupBy accepts PROCESS (default), TASK, USER, or GROUP. Optional dueDateFrom and dueDateTo must contain timestamp offsets and form an ordered range. Supports page and size, capped at 100.

curl "http://localhost:8080/admin/sla/metrics?groupBy=PROCESS&page=0&size=20" \
  -H "Authorization: Bearer $TOKEN"

The response is a page of SlaMetricRow records: status counts, completion rate within SLA, successful and failed escalations, and alert-to-completion sample count and average seconds. Rates and averages may be null when no samples exist.

Failed SLA actions can be retried through POST /incidents/{id}/retry by an administrator. Retry eligibility depends on the current task, process, and action state.