Instance view: "External Tasks" and "Jobs" tabs are unimplemented placeholders (no API resources)

[javahippie]

Summary

The "External Tasks" and "Jobs" tabs in the process instance view are unimplemented placeholders — they render static text only:

// src/pages/Processes.jsx
// TODO: create Jobs example for old Camunda apps
const InstanceJobsPlaceholder = () => <p>Jobs</p>;
// TODO: create External Apps example for old Camunda apps
const InstanceExternalTasksPlaceholder = () => <p>External Tasks</p>;

There is no /external-task and no /job resource anywhere in src/api/resources/ (and consequently no state signals), so this is a full-stack gap, not an empty fetch.

Expected (C7 Cockpit parity)

External Tasks tab — GET /external-task?processInstanceId={id}:
columns ID, Activity, Retries, Worker ID, Lock Expiration Time, Topic, Priority; error details drill-down via GET /external-task/{id}/errorDetails.

Jobs tab — GET /job?processInstanceId={id}:
columns ID, Due Date, Activity, Retries, plus failure log (GET /job/{id}/stacktrace).

Cockpit also offers actions (set retries / set priority for external tasks; set retries, suspend/activate, trigger due date for jobs). Suggest splitting those into a follow-up and starting with read-only tables — same pattern as the incidents tab (InstanceIncidents): resource file + signal in state.js + table component + i18n keys.

Note

Activity names require mapping activityId against the BPMN model (the diagram XML is already in state.api.process.definition.diagram), otherwise show the raw activity id.

[lambdaschmied2]

Both tabs were already implemented with tables and actions (the "placeholder" state predates this); the real remaining gap was the failure-log drill-downs, whose resources (external-task/{id}/errorDetails, job/{id}/stacktrace) existed but were never wired into the UI.

Addressed in commit 3259efa:

• External Tasks: added a Priority column and an Error Details dialog rendering GET /external-task/{id}/errorDetails.
• Jobs: added a Stacktrace dialog rendering GET /job/{id}/stacktrace.

Testing status — needs a proper dev-fixtures pass before closing.

• External Tasks Error Details drill-down was verified end-to-end against a manually-failed external task (dialog shows the real error details). ✅
• Jobs Stacktrace drill-down is not yet verified with live data: the dev-fixtures bot's failures surface as external-task errors/incidents, not failed jobs with stacktraces, so no job with an exceptionMessage was available to trigger the button. The code path mirrors the verified external-task drill-down and unit tests pass, but it should be exercised against a process that produces a genuinely failing (async) job in the dev-fixtures setup.

Leaving this open until the Jobs stacktrace path is confirmed with real fixture data. Also still open per the original issue's follow-up notes: activity-name mapping (activityId → BPMN model) and the extra Cockpit actions (set priority, trigger due date).