An embedded MQTT broker plus HTTPS batch ingest, schema-validated on every payload and locked down with scoped keys. No Telegraf, no database, no broker to stand up yourself.
Getting plant data off the floor and into something usable usually means stitching together a collector like Telegraf, a time-series database, and a separate MQTT broker, then writing the glue and the auth layer between all three. Every piece needs its own config, its own credentials, its own upgrade path, and its own 2 a.m. failure mode. Worse, nothing validates what actually arrives, so a firmware change on a sensor silently pushes garbage into your historian and you find out weeks later when a report looks wrong.
Sentrel ships the broker and the ingest path with the platform. Edge devices publish to the embedded MQTT broker, or batch readings over HTTPS when a persistent connection isn't practical. Every payload is checked against the typed schema you defined in plain English before it is ever stored, so the data that lands is the data you expect. Access is scoped per source: each device or integration gets its own API key or MQTT key with its own permissions, so a compromised gateway can't read or write anything beyond its lane.
The DIY stack gets you ingest after you've become the systems integrator for three open-source projects. Sentrel gets you validated, authenticated ingest the moment you define a schema, with one thing to operate instead of a pipeline. Because validation happens at the door, bad readings are caught at ingest instead of poisoning dashboards and reports downstream, and because keys are scoped per source, you can onboard or cut off a gateway without touching anything else. It is the same outcome the assembled stack promises, minus the assembly, the auth glue, and the maintenance tail.
No. Sentrel includes an embedded MQTT broker, so edge devices can publish directly without you deploying, securing, or licensing a separate broker. If a device can't hold a persistent connection, it can send readings in batches over HTTPS instead.
Every payload is validated against the typed schema you defined before it's stored. Readings that don't conform are rejected and surfaced at ingest rather than silently written, so a sensor firmware change or a malformed batch can't quietly corrupt your dashboards and reports.
Each device or integration gets its own scoped API key or MQTT key with its own permissions. Keys are issued and revoked independently, so you can onboard or cut off a single gateway without affecting any other source, and a compromised key can't reach beyond its assigned scope.
Describe what you want to watch, connect the gateway you already own, and be live this afternoon — no integrator, no proprietary hardware, no six-month project.