Skip to main content

OpCon RPA Task job

In briefAn OpCon RPA Task job starts one Robot, Web Macro or Scan Document task on an OpCon RPA agent, waits for it to finish, and shows the agent's logs as the job output. The status the agent reports decides the outcome. The task and version lists are read live from the agent. The agent is reached through a relay, but it is not an LSAM.

WhereRegister the agent in Agents › Agents with New Agent, choosing OpCon RPA as the agent type and entering its GUID API Token. Build the job in Design › Workflows: open a workflow, open the job in the Job Editor (open its context menu on the canvas › Edit Job), choose OpCon RPA Task as the job type, pick the agent under Agent Assignment, and then set Sub-type, Task and Version under Job Parameters.

WhenUse an OpCon RPA Task job when:

  • a workflow has to run a robot, a browser macro or a document scan that already exists on an RPA agent;
  • later jobs depend on whether that task completed;
  • you want the task's own logs kept with the job run.

How

  1. Use a relay that supports OpCon RPA, and register the agent with its host, HTTPS Port and GUID API Token.
  2. In the job, pick the agent first, then the Sub-type, because the Task list depends on both.
  3. Choose the Task, and a Version or Use latest.
  4. Save the workflow.
Notes
  • Completed is success. Skipped, Failed and Aborted are failures. There is no exit-criteria setting.
  • The job sends the job's and the workflow's instance properties to the task, as stored and not resolved. Encrypted properties are left out.
  • Use latest is resolved each time the job starts, not when you save.
  • If a lookup fails, the Task and Version fields still accept a typed value.
Watch out
  • An agent with no stored API token reports DOWN, and no job can run on it.
  • If the job's and workflow's properties together are over 64 KiB, the job fails and nothing runs.
  • A restarted job always starts a new process on the agent. It does not resume the earlier one.
  • After Relay restarted before the RPA process id was known, check the agent before you run the job again. A robot may still be running.