Inbound webhooks

Inbound webhooks let external systems create Manage365 alerts via HTTP. Each webhook has a unique URL + bearer secret; the alert flows through the same channels, escalation rules, and PSA ticket automation as any other alert.

Creating a webhook

Settings → Notifications → Inbound webhooks → New webhook. Give it a name (e.g. "n8n server-down detector") and save. You'll see the POST URL + bearer secret exactly once.

Sending an event

curl -X POST https://api.manage365.com.au/api/v1/inbound/abc123… \
  -H "Authorization: Bearer m365_hook_…" \
  -H "Content-Type: application/json" \
  -d '{
    "severity": "high",
    "title": "Server unreachable",
    "description": "Monitoring probe failed 3 times in a row.",
    "customerTenantDomain": "client.onmicrosoft.com",
    "dedupeKey": "prtg:server-123"
  }'

Payload fields

Field	Required	Notes
`title`	yes	Short headline — shows on alert rows and Teams/Slack
`description`	yes	Longer body. Plain text.
`severity`	no	`low` \| `medium` \| `high` \| `critical`. Defaults to medium.
`category`	no	`system` (default) \| `security` \| `compliance` \| `licence` \| `drift` \| `billing`
`customerTenantId`	no	Manage365 UUID. Use this if you have it.
`customerTenantDomain`	no	M365 default domain (`client.onmicrosoft.com`). Nicer for n8n flows — no UUIDs to look up.
`dedupeKey`	no	If you send the same key twice, the second call updates the existing alert instead of creating a new one.
`metadata`	no	Arbitrary JSON object — surfaced in alert details.

Response

202 Accepted with { "ok": true, "alertId": "…" } on success. The alert is queued immediately — notification channel fan-out and PSA escalation happen async.

Failure modes

401 — wrong bearer secret, or the webhook has been deleted.
404 — slug doesn't exist, or the webhook is disabled.
400 — missing title or description, or customerTenantDomain not found on your MSP.

Monitoring

Each hook tracks success + failure counts + last fired/last error on the card. Broken integrations show up in red with the last error message visible — debugging doesn't need server logs.