Telemetry was lost on the way
Weik.io 2026.2The hub lost telemetry in the last 15 minutes. The detail has each cause and the action that fixes it.
| Kind | Severity | Who sees it |
|---|---|---|
telemetry- |
Warning | Platform administrators, in a list that isn’t narrowed to an environment or a project |
What it means#
The hub lost some of the telemetry it was sent in the last 15 minutes, so the charts and searches of that time miss part of what happened.
What raises it#
- Telemetry the hub couldn’t store or keep on its disk for later
- Metrics that went over the metric store’s limit on new series
The detail says what was lost and has an action for each cause. The hub doesn’t store a batch of telemetry with more than 16,384 entries or more bytes than the installation’s telemetryBatchMaxBytes, which is 8 MiB by default. For such batches, the detail also lists their senders. An installation set up without the hub’s metric store doesn’t raise this problem.
A batch that a hub server answered with 503 while it was busy with other batches doesn’t count as lost, because its sender sends it again.
What to check#
See the stores and what Weik.io received:
$ weikio platform healthOn a hub server, in the stack’s folder, check the stores:
$ docker compose run --rm infra weikio-backend doctorHow to fix it#
Follow the action in the detail for each cause:
- When a store didn’t answer, start it again with
docker compose up -din the stack’s folder. - When metrics went over the metric store’s limit on new series, look first for an integration’s custom counter with a tag that changes with every message, such as an order id.