Skip to main content

Notifications overview

Receive real-time updates when your users record new health data — no polling required. Withings POSTs to your callback URL the moment new measurements arrive, then your handler fetches the data via the Data API.

Why webhooks

The standard Withings API enforces a 120 requests-per-minute limit across your application (see API plan details). Polling for fresh data hits this ceiling quickly and adds latency you can't tune.

The notification system inverts the flow: Withings tells your backend exactly when new data is ready, so you fetch only what you need, exactly when it matters.

If your use case demands a higher request rate or stricter SLA guarantees, the Enterprise plan unlocks higher limits and dedicated support.

How it works

  1. 1

    User measures

    A user takes a measurement on their Withings device.

  2. 2

    Withings detects

    Our servers identify new data for the relevant appli category.

  3. 3

    Webhook POST

    Withings sends an application/x-www-form-urlencoded POST to your callback URL.

  4. 4

    Your handler responds

    Your endpoint returns HTTP 2xx within a few seconds.

  5. 5

Reliability

Two layers of safety so a hiccup in your endpoint doesn't cost you data.

Per notification — 5 retry cycles over ~5 hours

If your endpoint doesn't return HTTP < 400, Withings retries the same notification on this schedule:

  1. T+0Initial
  2. +10 sRetry 1
  3. +1 minRetry 2
  4. +1 hRetry 3
  5. +4 hRetry 4

Each cycle does 2 HTTP attempts (10 attempts total). Later retries use random jitter to spread load.

Per subscription — failure escalation

When notifications keep failing despite the retries above, your subscription escalates:

  • 4 failed notifications in a rowA warning email is sent for each (max 4 emails).
  • 20 days of unresolved failuresSubscription is auto-cancelled, with a final notice email.

Missed notifications are not redelivered. Once your endpoint is back up, backfill via the Data API using lastupdate.

Real-time, but not guaranteed

Typical delivery delay is under 2 minutes, but it can occasionally be longer. Withings does not commit to a real-time SLA on the standard plan. Need stricter guarantees? See the Enterprise plan.