Sealing deadline, paused telemetry and lost seals

What you're seeing

One of these on the Home page, the Brief or the Findings page, or on an emitter in Setup:

  • Sealing deadline approaching (data_health.seal_deadline): Medium, or High within 14 days. "This emitter still sends batches without its seal. From 2027-03-01 (45 days from now) its telemetry will be paused until its SDK is upgraded."
  • Telemetry paused (data_health.telemetry_paused): High. "Telemetry paused: 12 batches refused since 2027-03-02 because they lack the emitter's seal. Nothing from this emitter is being stored until it is upgraded."
  • Batches lost their seal (data_health.unsealed_batches): Medium. "2 batches since batch 41 arrived without the emitter's seal, although it sealed before."

What it means

Sealing is the emitter signing the hash of every record in each batch, so an auditor can check that the emitter itself vouched for the records. See The emitter proof.

  • Approaching means Provantix has sent the notice, so the date from which unsealed batches are refused is set, and this emitter has not moved to sealing. Your admins get the notice by email, then reminders 30 and 7 days before the date. Nothing is lost yet.
  • Paused means the date has passed and this emitter's batches are being refused. The emitter queues them on its host and retries. Nothing it queued is lost, but nothing is stored until it is upgraded.
  • Lost their seal means an emitter that used to seal now sends unsealed batches. The records are still accepted, but the emitter no longer vouches for them, so an auditor loses the emitter proof for everything sent from then on.

Pilot and trial sources are exempt from the deadline and never appear under the first two.

What to do

  1. Upgrade the SDK to 0.2.0 on the emitter named in the finding and restart the service. Sealing is on by default.
  2. Nothing else changes. The emitter id, the key and the endpoint stay the same.
  3. Batches already queued on the host are sealed when it sends them, with the same records, batch number and creation time. A paused emitter loses nothing it stored up. Refused batches show in Connectors under the rejection code seal_required.
  4. The finding resolves itself when a sealed batch is accepted.

If it says lost their seal

  • Check which SDK version the service now runs. A rollback or a redeploy from an old image is the usual cause.
  • Check that no second process sends under the same emitter id with an older SDK.
  • Check that seal: false was not set.

Why a refusal says 426

The platform answers an unsealed batch after the deadline with 426 seal_required, the standard "upgrade required" status, and a message that names the fix. The SDK treats it like any refused delivery: it keeps the batch spooled and retries, and a version that seals re-signs the queue before sending. See the batches API.