Classify stops
Required scope: stop_classifications.update. Related: reports.read to find the stops to classify with POST /v1/reports/stops; stop_catalogs.read to list the causes and components each machine accepts with GET /v1/stops/classification-options.
Classify stops: set their general cause, specific cause, component and description. The body is a JSON array of up to 1,000 rows, each its own classification.
Identifying a stop. A stop is identified by its machine and the time it started, not by an
id: stop ids can change when a shift is recalculated. Take machine.id and stop.startTime
from POST /v1/reports/stops. Open stops, still running, can be classified too.
Choosing a classification. Each machine accepts only the causes of its loss tree and its
own components, listed by GET /v1/stops/classification-options. A row sets at least a
generalCauseId or a description; specificCauseId needs a generalCauseId.
A row replaces the stop's classification. Whatever a row leaves out is cleared: sending
only generalCauseId removes the stop's specific cause, component and description. Sending
the same rows again is harmless, so a request that timed out can be retried as is. A cause
set by an operator in Pulsar is overwritten too.
Every row comes back with a status, and only classified rows changed anything:
classified: the stop now has this classification.stopNotFound: machinemachineIdhas no stop that started atstartTime, or it is not one of your machines.duplicateStop: another row in the same request names the same stop, so neither was applied.generalCauseNotAvailable: the general cause is not in the machine's loss tree.specificCauseNotAvailable: the specific cause does not go under the general cause on this machine.componentNotAvailable: the component is not assigned to the machine.
The schema also lists missingClassification and missingGeneralCause. This endpoint rejects
such rows with 422 before they reach Pulsar, so you will not receive them.
Result rows echo your rows in request order, with startTime normalized to UTC; match them by
position rather than by comparing startTime strings.
The classified rows of a request are saved together. Metrics that depend on the stops'
causes are recalculated in the background, so reports may take a moment to show the change.
Request Body
application/json
TypeScript Definitions
Use the request body type in TypeScript.
A JSON array. Each item has these fields:
Response Body
application/json
application/json
application/json
application/json
application/json
curl -X PUT "https://example.com/v1/stops/classifications" \ -H "Content-Type: application/json" \ -d '[ { "machineId": 12, "startTime": "2026-10-12T10:00:00Z" } ]'[ { "machineId": 12, "startTime": "2026-10-12T10:00:00Z", "generalCauseId": "d56aece5-a92a-462e-b275-919e15446bba", "specificCauseId": "f89b5866-2d26-45a1-8cfc-8295557594d8", "componentId": "3e8e81cf-7521-4ef9-b6c9-40074aa764c5", "description": "Bearing seized on the main motor", "status": "classified" }]List how each machine's stops can be classified GET
List, per machine, the classifications PUT /v1/stops/classifications accepts for its stops: * generalCauses: the general causes of the machine's loss tree, each with the specificCauses that can go under it. * components: the components assigned to the machine. A machine without a loss tree has no general causes; its stops can only be given a description. Narrow the list with machineId, repeated for several machines: ` GET /v1/stops/classification-options?machineId=12&machineId=13 `
Stop created Webhook
Sent when Pulsar records a new machine stop, whether or not it's classified. The cause fields are null until someone classifies it.