Sign inStart your trial

Troubleshooting

Common problems and how to fix them

When something stops working, this page covers the most common cases. For general questions about how Rills works, see the docs home . For anything not covered here, email support@rills.ai or chat with Wade in the app.

#A workflow run failed

Failed runs are flagged in the workflow’s Activity tab with a red status. Click the failed run to see which step failed and why.

Most common causes:

  • Disconnected integration: an OAuth token expired or was revoked. The integration page shows a warning. Reconnect it.
  • Missing config: a node references a field or variable that isn’t set. Open the node and check for empty required fields.
  • Bad input: the trigger payload or an earlier step’s output doesn’t match what the next step expects. Check the run detail view to see the actual data.
  • Timeout: a step (usually custom code) ran past its timeout. Increase the timeout or break the work into smaller steps.

If you’ve fixed the cause, re-run the workflow from the failed run’s detail page.

#A custom code step crashed

Custom code shows one of four error types:

  • Out of memory: your code allocated more than the sandbox allows. Process data in chunks or filter before loading.
  • Timeout: execution went past the configured timeout (max 300 seconds). Increase it or split the work.
  • Crash: your code threw an unhandled exception. Check the stderr output in the run detail view.
  • Sandbox failure: rare infrastructure issue. Retry; contact support if it persists.

For all of these, the run detail view shows captured stderr, start there.

#A webhook isn’t triggering

Webhook calls return HTTP status codes that tell you what went wrong:

  • 400: the body isn’t valid JSON or the URL is malformed. Check your request body and Content-Type: application/json header.
  • 401: auth failed. For API key, check the header name and value. For HMAC, check the signing secret and algorithm.
  • 404: no webhook with that ID. Copy the current URL from the workflow trigger config, it may have changed.
  • 429: rate limited. Honor the Retry-After header.
  • 503: webhook disabled. Re-enable it in the workflow’s trigger settings.

To test a webhook quickly: use curl against the URL with a sample payload and check the response.

#An integration disconnected

When an OAuth token expires or is revoked, every workflow using that integration will fail. The integration page flags broken connections.

Common reasons:

  • Token expired and refresh failed (refresh tokens age out, especially after long inactivity)
  • You removed Rills from the service’s authorized apps page
  • The service updated its API and now requires new scopes

Fix: open Integrations, find the broken one, click Reconnect, and re-authorize. Workflow configurations are preserved across reconnects, only the credential changes. Re-run any failed runs that depended on it.

#Credits ran out

When you exceed your included quota and hit your spending cap, workflows that need to spend credits pause. Three options:

  • Wait: credits reset at the start of the next billing cycle.
  • Raise the spending cap: allows overage billing at your plan’s per-unit rate.
  • Upgrade the plan: more included credits and lower overage rates.

Open Settings → Billing to see usage and adjust.

#Usage is higher than expected

Common culprits:

  • Scheduled workflows with frequent trigger checks (each check costs Workflow Credits even if the workflow doesn’t proceed)
  • AI nodes with large prompts or oversized output (more tokens = more AI Credits)
  • Retry loops on a failing step that keeps consuming credits

Review Settings → Usage for a per-workflow breakdown. The biggest spenders are usually the suspects.

#A proposal expired before I could approve it

Approval gates have a timeout, by default 24 hours, configurable per gate (max 30 days). Expired proposals trigger the timeout action you configured: Approve, Reject, Escalate, or Fail. Default is Fail so stale proposals don’t run after the moment has passed.

If your team is missing approvals, raise the timeout, set up push notifications , or assign proposals to a specific team member who’s monitoring.

#Something else

  • Help in-app: chat with Wade for debugging or Brooke for general triage.
  • Email: support@rills.ai with your workspace name, the workflow name, the run ID if relevant, and what you’ve already tried.