It is possible! You can find more details on the Cribl API Documentation, but Ive listed the general steps below.

You will need to obtain the JSON configuration of the pre-configured collector. NOTE: If the collector doesnt already exist you will need to create it before running these steps!

GET /api/v1/m/<worker-group-name>/lib/jobs returns the JSON of the collector configurations. Find the one with the corresponding id field.

{  "items": [  {  "type": "collection",  "ttl": "4h",  "removeFields": [],  "resumeOnBoot": false,  "schedule": {},  "collector": {  "conf": {  "discovery": {  "discoverType": "http",  "discoverMethod": "get",  "itemList": [],  "discoverDataField": "entry",  "discoverUrl": "`https://1.2.3.4:8089/services/search/jobs`",  "discoverRequestParams": [  {  "name": "output_mode",  "value": "`json`"  },  {  "name": "search",  "value": "`\"search index=_internal\"`"  }  ]  },  "collectMethod": "get",  "pagination": {  "type": "none"  },  "authentication": "login",  "loginUrl": "`https://1.2.3.4:8089/services/auth/login?output_mode=json`",  "loginBody": "`username=${username}&password=${password}`",  "tokenRespAttribute": "sessionKey",  "authHeaderExpr": "`Splunk ${token}`",  "username": "admin",  "password": "redacted",  "collectUrl": "`${id}/results`",  "collectRequestHeaders": [],  "collectRequestParams": [  {  "name": "output_mode",  "value": "`json`"  }  ]  },  "destructive": false,  "type": "rest"  },  "input": {  "type": "collection",  "staleChannelFlushMs": 10000,  "sendToRoutes": true,  "preprocess": {  "disabled": true  },  "throttleRatePerSec": "0",  "breakerRulesets": [  "splunk_test"  ]  },  "id": "splunk",  "history": []  },  ...  ]}

Use this data to POST back to /api/v1/m/<worker-group-name>/jobs with an added run field with the configuration.

For example, I want to run this collector in preview mode:

{  "type": "collection",  "ttl": "4h",  "removeFields": [],  "resumeOnBoot": false,  "schedule": {},  "collector": {  "conf": {  "discovery": {  "discoverType": "http",  "discoverMethod": "get",  "itemList": [],  "discoverDataField": "entry",  "discoverUrl": "`https://1.2.3.4:8089/services/search/jobs`",  "discoverRequestParams": [  {  "name": "output_mode",  "value": "`json`"  },  {  "name": "search",  "value": "`\"search index=_internal\"`"  }  ]  },  "collectMethod": "get",  "pagination": {  "type": "none"  },  "authentication": "login",  "loginUrl": "`https://1.2.3.4:8089/services/auth/login?output_mode=json`",  "loginBody": "`username=${username}&password=${password}`",  "tokenRespAttribute": "sessionKey",  "authHeaderExpr": "`Splunk ${token}`",  "username": "admin",  "password": "redacted",  "collectUrl": "`${id}/results`",  "collectRequestHeaders": [],  "collectRequestParams": [  {  "name": "output_mode",  "value": "`json`"  }  ]  },  "destructive": false,  "type": "rest"  },  "input": {  "type": "collection",  "staleChannelFlushMs": 10000,  "sendToRoutes": true,  "preprocess": {  "disabled": true  },  "throttleRatePerSec": "0",  "breakerRulesets": [  "splunk_test"  ]  },  "id": "splunk",  "history": [],  "run": {  "rescheduleDroppedTasks": true,  "maxTaskReschedule": 1,  "logLevel": "info",  "jobTimeout": "0",  "mode": "preview",  "timeRangeType": "relative",  "expression": "true",  "minTaskSize": "1MB",  "maxTaskSize": "10MB",  "capture": {  "duration": 60,  "maxEvents": 100,  "level": "0"  }  }}

This returns a JSON response with a Job id: {"items":["1621367040.54.adhoc.splunk"],"count":1}

You can then query the jobs endpoint to get the status of the job.

GET /api/v1/m/<worker-group-name>/jobs/1621367040.54.adhoc.splunk

Which provides a JSON response (check the status.state field for more information):

{  "items": [  {  "id": "1621367040.54.adhoc.splunk",  "args": {  "type": "collection",  "ttl": "60s",  "removeFields": [],  "resumeOnBoot": false,  "schedule": {},  "collector": {  "conf": {  "discovery": {  "discoverType": "http",  "discoverMethod": "get",  "itemList": [],  "discoverDataField": "entry",  "discoverUrl": "`https://1.2.3.4:8089/services/search/jobs`",  "discoverRequestParams": [  {  "name": "output_mode",  "value": "`json`"  },  {  "name": "search",  "value": "`\"search index=_internal\"`"  }  ]  },  "collectMethod": "get",  "pagination": {  "type": "none"  },  "authentication": "login",  "loginUrl": "`https://1.2.3.4:8089/services/auth/login?output_mode=json`",  "loginBody": "`username=${username}&password=${password}`",  "tokenRespAttribute": "sessionKey",  "authHeaderExpr": "`Splunk ${token}`",  "username": "admin",  "password": "redacted",  "collectUrl": "`${id}/results`",  "collectRequestHeaders": [],  "collectRequestParams": [  {  "name": "output_mode",  "value": "`json`"  }  ],  "filter": "(true)",  "discoverToRoutes": false,  "collectorId": "splunk",  "removeFields": []  },  "destructive": false,  "type": "rest"  },  "input": {  "type": "collection",  "staleChannelFlushMs": 10000,  "sendToRoutes": false,  "preprocess": {  "disabled": true  },  "throttleRatePerSec": "0",  "breakerRulesets": [  "splunk_test"  ],  "output": "devnull",  "pipeline": "passthru",  "filter": "(true)"  },  "id": "splunk",  "history": [],  "run": {  "rescheduleDroppedTasks": true,  "maxTaskReschedule": 1,  "logLevel": "info",  "jobTimeout": "0",  "mode": "preview",  "timeRangeType": "relative",  "expression": "true",  "minTaskSize": "1MB",  "maxTaskSize": "10MB",  "capture": {  "duration": 60,  "maxEvents": 100,  "level": "0"  },  "type": "adhoc",  "taskHeartbeatPeriod": 60  },  "initialState": 3,  "groupId": "default"  },  "status": {  "state": "finished"  },  "stats": {  "tasks": {  "finished": 1,  "failed": 0,  "cancelled": 0,  "orphaned": 0,  "inFlight": 0,  "count": 1,  "totalExecutionTime": 80,  "minExecutionTime": 80,  "maxExecutionTime": 80  },  "discoveryComplete": 1,  "state": {  "initializing": 1621367040902,  "paused": 1621367040903,  "pending": 1621367041736,  "running": 1621367041738,  "finished": 1621367041852  }  },  "keep": false  }  ],  "count": 1}

Isn't there a better way of doing this? Sending a huge body in an API request is just asking for errors. I've been trying to use this setup to run some script collectors via API and having no luck. I also have around a dozen script collectors I want to kick off programmatically via the API rather than put on a schedule. Sending a huge ungainly API command for each one is a bit untenable. If there isn't a better way of doing this I suggest the addition of a RunJob API command which has a minimal set of inputs, and its generally beter to have those inputs be parameters rather than prone to error JSON bodies.

Current error being worked, though I might drop this and find a better way:

{    "status": "error",    "message": "invalid config jobs: n{\"keyword\":\"required\",\"dataPath\":\"/ohkgZv\",\"schemaPath\":\"#/definitions/collection/required\",\"params\":{\"missingProperty\":\"collector\"},\"message\":\"should have required property 'collector'\"},{\"keyword\":\"if\",\"dataPath\":\"/ohkgZv\",\"schemaPath\":\"#/patternProperties/.*/if\",\"params\":{\"failingKeyword\":\"then\"},\"message\":\"should match \\\"then\\\" schema\"}]"}