Workflow stuck

A workflow run stays queued, pauses on one node, or never reaches a completed state.

Likely causes

Background job processing is down, so queued steps never start.
A node is waiting on an external response (an integration call, a webhook, or a human approval).
A condition routed the run down a path with no further steps.
The workspace ran out of credits during a step that calls a model.
A node errored and the run halted instead of continuing.

Open the execution history
Go to Workflow execution history and open the run. Each node shows its status and output, so you can see exactly where the run stopped.
Confirm background jobs are healthy
Workflows execute through the Inngest background job processor. If a run is queued but never starts, test the connection in Settings → Diagnostics. An unhealthy connection blocks all workflow execution.
Inspect the stalled node
A node waiting on an external call (an integration or webhook) or a human approval will sit in a pending state by design. Confirm the upstream system responded, or that the approver acted. For routing problems, review your conditional logic.
Check credits for model steps
If the run stalls on a step that calls an agent or model, the workspace may be out of credits. Verify at Settings → Usage & Billing.
Fix and re-run
Correct the failing node or condition, then re-run the workflow. Use Testing a workflow to validate the path before relying on it in production.

Execution history — /automation/workflows/execution-history for per-node status.
Diagnostics — /settings/diagnostics to confirm background job processing is up.
Usage & Billing — /settings/usage-billing for balance.