Usage alerts
What fires
Two alerts on the monthly send allowance, at 80% and at 100%:
| to the workspace’s alert recipient | |
| Webhook | tenant.usage_threshold |
| Frequency | once per level per billing period |
So a workspace that crosses 80%, keeps sending, and crosses 100% gets two alerts that period and no more. Crossing 80% again next period alerts again.
Subscribe to the event the same way as any other:
{ "url": "https://yours.example/hooks", "events": ["tenant.usage_threshold"] }
What does not fire
Stated so you do not wait for something that is not coming:
No alert on the per-minute cap. It is hit and recovered from constantly
by design, and alerting on it would be noise. You see it as a 429.
No alert for api-mode workspaces on the monthly allowance, because they have none: their plan is unit-based, and exhaustion is a units balance reaching zero rather than a monthly pool filling up. The evaluation skips zero-allowance plans entirely.
No alert at an arbitrary threshold you choose. 80 and 100 are the two,
and they are not configurable. If you want 50%, poll
GET /v1/usage and decide for yourself.
No alert on storage reaching the plan’s allocation through this path. Mailbox storage has its own warnings at 80, 90 and 100 per cent, which are about a mailbox filling up rather than about your API quota.
Reputation is the other thing worth watching
A high bounce rate lowers your rate cap, and that is a quota change you did
not make. It raises its own events rather than a usage threshold:
tenant.throttled when the cap drops.
If you subscribe to one platform event, subscribe to that one. A 429 from a
throttle looks exactly like a 429 from your ordinary cap, and only the
event or GET /v1/reputation distinguishes
them.