#Hub activity

Use the view for the path you configured: a Channel Route or an Automation event.

#Where to look

In the app, Channels → Activity lists messages evaluated against Routes. Filter by Connection, Route and outcome; open an event for its reason and recovery actions. A test reply checks outbound delivery only and does not appear as an inbound Route run.

Automations → your Automation → Runs shows Workflow results. The Hub dashboard's Activity lists organization trigger runs; Connections → Known unrouted events records accepted provider events that matched no Workflow:

Reason	Meaning
no_workflow_route	No Workflow route is configured for this event.
no_trigger_for_source	No configured trigger handles the source.
trigger_filters_rejected	The source matched, but filters rejected the event.
configuration_unavailable	A relevant configuration or Connection could not be resolved.

#Nothing happened when I mentioned the bot

1. Which path is configured? For a Route, check its Rules: Where, Who and message conditions. For a direct event, check from_users, resource and content filters. GitHub uses logins; prefer stable Slack/Discord user IDs. See Channels or Triggers.
2. Did Hub receive it? Look in Channel activity, Automation Runs or Known unrouted events. If absent, check Connection status and the provider delivery log. GitHub exposes App → Advanced → Recent Deliveries; Slack's direct-event path uses Event Subscriptions.
3. Is the intended revision active? A failed save or sync leaves the old revision serving. Check configuration and errors before recreating anything.
4. Is the Host connected? Confirm the target and its Hub relationship has hub.execute. A provider event dispatch can fail with daemon_not_connected.
5. Did it stop early? Compare the run with its step's idle_timeout, max_runtime, and Workflow deadline. Inspect author-access rejection separately from provider startup failure.
6. Did it run without replying? Check both the input's reply settings and the responding step's output grant. The prompt must name hub.reply and hub.finish_execution. See Workflow instructions.

#Sync failures

Repository-backed Configuration shows the last sync attempt. Fetch failed means the file is absent at that commit or the App cannot read it. Invalid means YAML or resource resolution failed. Superseded push ignored means a newer commit already moved the branch head.

Check renamed Hosts, unavailable Connection slugs and repository installation coverage. See Repository bundles.

#Recovery and retries

Channels have a durable Ingress queue. Retryable failures and busy Routes can wait for another attempt; retries are bounded. Inspect waiting work and Dead letters before sending a duplicate request. An Organization Admin or the Connection's Admin can resubmit its dead letters after fixing the cause.

This queue does not promise replay for every external Automation event. A failed direct-event run may need a fresh invocation after the Host returns. Check its final state and any earlier side effects first.