Each organisation can shape incidents to its own way of working without changing OpsCommand's base. The eleven statuses, the lifecycle between them and the eleven categories stay as they are. On top of them an organisation adds:
| What | Where it lives | Rules |
|---|---|---|
| Its own fields | ops.custom_fields, values in ops.incidents.custom | Types: text, number, date, choice, yes or no. A field keeps its type. Choices are added, never taken away while in use. A required field is required when an incident is opened. Archived, not deleted: its values stay on the incidents and it takes no new value. Only defined fields are accepted. |
| Subcategories | ops.subcategories, ops.incidents.subcategory | Each belongs to one of the eleven categories. Choosing one sets the incident's category. |
| Sub-statuses | ops.substatuses, ops.incidents.substatus | Each belongs to one of the eleven statuses. It can be set only while the incident has that status, and is cleared when the incident moves on. |
| Extra levels | ops.levels (ops.add_level) | A severity or priority at a free rank, 1 the most serious. A rank is taken once. Added through the API (POST /v1/organisations/{org}/workflow/levels). |
| Automation rules | ops.automation_rules, runs in ops.automation_runs | See below. |
Automation rules
A rule is: when an incident is opened, or its status changes, or its severity changes; and it matches the conditions (lists of services, categories, severities or statuses; an empty condition matches everything); then it acts, in order:
- assign to a team. Only through the lifecycle: if the incident's status cannot move to assigned, the action is skipped and the run says why.
- notify a team: every active member gets an escalation notification.
- escalate to a severity. Only ever raises it, never lowers it.
- set priority.
Each rule names only this organisation's records; the database checks every team, service and level when the rule is saved. Rules run in the database (ops.run_automation, a trigger after an incident is inserted or its status or severity changes), as the organisation and not as the person who made the change. A rule's own changes do not set off other rules, so rules cannot loop.
Every run is logged: the incident, the event, and each action with done or skipped and why. The incident's timeline shows a note "Automation: rule name", and the assignment event names the rule, with no person as its actor.
Who does what
- Every member reads the definitions, since they shape the incident form.
- Only
organisation.managedefines fields, subcategories, sub-statuses, levels and rules, and reads rules and their runs. Every definition is in the audit chain. - It needs the plan feature
advanced_workflows(Professional and Enterprise; not Starter). On a plan without it, nothing new is defined and no rule runs; what was defined stays, and incidents keep their values.
Where
- Database: migration
0050_workflows. - API:
GET /v1/organisations/{org}/workflow;PUT …/workflow/fields/{key},…/workflow/subcategories/{code},…/workflow/substatuses/{code};POST …/workflow/levels;POST …/workflow/rules,PUT …/workflow/rules/{id};GET …/workflow/rules/{id}/runs. Incidents carrycustom,subcategoryandsubstatus. - Web: Administration → Workflow; the organisation's fields on the incident forms and in the incident's details.