Telemetry ingest that survives bad networks
Devices push metrics and events over a single long-lived HTTPS stream. Metricsink buffers, batches and forwards to your warehouse — with backpressure, not data loss.
POST https://api.metricsink.dev/v1/ingest
Authorization: Bearer ms_live_…
Content-Type: application/x-ndjson
{"ts":1757980020,"dev":"gw-0417","temp_c":41.2,"rssi":-67}
{"ts":1757980021,"dev":"gw-0417","temp_c":41.3,"rssi":-66}
What you get
- Streaming ingest over HTTP/2 with keep-alive; reconnect-safe cursors
- Per-device buffering up to 24 h when the sink is unavailable
- Delivery to S3, ClickHouse, Kafka or a plain webhook
- Regional endpoints in eu-north and eu-central
Beta. Ingest API is stable; delivery adapters may change. See docs.