Run center, logs, graphs and stop

Entry and workflow

Run center → trigger/instance details. All three roles can read run summaries. Stop, pause, resume and retry require ADMIN or OPERATOR and an instance state that allows the action. Logs and input/output details require ADMIN. See the role matrix for all actions.

Configuration and actions

Run information

Display Meaning Shown when
Trigger ID Identifies a request. Acceptance does not imply an instance exists or execution succeeded. Always
Root instance ID Identifies a root run; retries reuse it. Root instance has been created
Task definition ID Task definition snapshot used by this run, not the current draft. Task instance summary
Job definition ID Job definition snapshot and fixed references used by this run. Job instance summary
Started Actual execution start time; may be absent while waiting. Always
Ended Time when execution reached a terminal state. Always
Duration (ms) Recorded elapsed time in milliseconds. Always
Status Distinguish waiting, running, paused and terminal states. SUCCESS still requires business verification. Always
Logs Read context for the selected root or node; the view contains up to the latest 5000 lines. ADMIN; logs are readable for the selected instance
Input variables Inputs used by this run, not later template defaults. ADMIN; Variables tab
Output variables Selected-node outputs. Redaction, truncation and unavailable notices mean the preview is not complete raw data. ADMIN; Variables tab
Instance tree Relationship of root, node and child instances; selection changes the information scope. Always

Instance controls

Action Roles Available when Result
Pause ADMIN / OPERATOR Root instance is RUNNING Submit a pause command, effective at execution checkpoints.
Resume ADMIN / OPERATOR Root instance is PAUSED Continue the same root run.
Stop ADMIN / OPERATOR Root instance is nonterminal, such as waiting, running, paused or waiting for an event Request stop and verify final status. Already completed external writes are not undone.
Retry ADMIN / OPERATOR Root instance is FAILED or TIMED_OUT Reuse the original version and root instance. The server selects failed job branches; external operations may repeat.

Use and verify

docs-hello should show SUCCESS and [STDOUT] hello-ijob; inspect output variable rows. The workflow graph should complete hello→check. Log pagination/truncation does not limit actual execution. Supported pause acts at step boundaries; current external work may finish and pause time still counts toward timeout.

Real page examples

Actual logs filtered through the UI to the hello output; clear search to view other loaded lines.

Actual logs filtered through the UI to the hello output; clear search to view other loaded lines.

Both nodes of the actual workflow run succeeded.

Both nodes of the actual workflow run succeeded.

Documentation version 1.0.0. Replace sample paths, resources and ports with those of your own environment.