Health and troubleshooting

When to use it

Check source health and the named agent's evidence when expected activity does not arrive or a receiver stops accepting notifications.

Set it up

1. Open Agents & outputs and inspect the agent's active settings, test evidence, message ID, and notification ID. Check the selected connections under Sources.
2. For Reconnect needed, authorize the source again. For a removed Teams app, add it back and reconnect the installation.
3. Use get_notifier_diagnostics with the address's notifierId and get_destination_diagnostics with receiver.destinationId to inspect holds, delivery failures, and incidents.
4. If no message entered the inbox, check capture choices and conditions with fresh samples. If a message was retained but not delivered, check notification timing and receiver polling or webhook history.
5. Refresh the evidence. Correlate message and signal IDs at the receiver when delivery succeeded but work did not.

Important limits

• Silence alone is not an outage. A successful mailbox scan is not proof that an email matched.
• Degraded means at least one selected account needs reconnection while available sources can continue.
• A connection incident stays unresolved until reconnection. A destination incident clears when that destination accepts a wakeup again.
• Delivery health measures the destination’s ability to receive notifications, not matching accuracy or completed agent work.
• Reconnection does not replay missed email.

Troubleshooting

• If health evidence is unavailable, check again or reload rather than assuming the last displayed status is current.
• If monitoring and delivery are healthy, inspect the receiving agent’s logs, task instructions, separate mailbox access, and polling schedule.