ChirpStack to Grafana: Visualize LoRaWAN Data

Grafana is a dashboard and alerting layer. It does not run a LoRaWAN network server and it does not receive uplinks. It queries a datasource on a schedule and draws the result. So the integration is really two hops: ChirpStack writes decoded uplinks into a time-series store, and Grafana reads from that store.

There are two common stores:

• InfluxDB is the shortest path, because ChirpStack ships a built-in InfluxDB integration. No extra code.
• PostgreSQL with TimescaleDB suits teams already on PostgreSQL who want SQL, joins and relational device metadata next to the measurements.

This guide focuses on the InfluxDB path and then shows the MQTT to Telegraf alternative.

In the ChirpStack application, open Integrations and add InfluxDB. ChirpStack supports both InfluxDB versions:

• InfluxDB v2: API endpoint, organization, bucket and an API token.
• InfluxDB v1: endpoint, database, username and password.

For a v2 setup the values look like this:

Version:          InfluxDB v2
API endpoint:     https://influx.example.eu/api/v2/write
Organization:     merkaio
Bucket:           chirpstack
Token:            <write-token-from-influxdb>

Once enabled, every decoded uplink is written automatically. ChirpStack stores the device fields as measurements and adds tags such as the DevEUI and application, so you can filter per device in Grafana later.

Create a panel and query the field you want. With Flux, filter by measurement and field and group by the device tag so each device gets its own series:

from(bucket: "chirpstack")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r._measurement == "device_frmpayload_data_temperature")
  |> filter(fn: (r) => r._field == "value")
  |> aggregateWindow(every: v.windowPeriod, fn: mean, createEmpty: false)
  |> group(columns: ["dev_eui"])

The exact measurement name depends on the integration version and the decoded field name; use the Grafana query inspector or the InfluxDB UI to read the real names from your data. From here you add humidity, battery and RSSI panels the same way and wire Grafana alert rules on top.

ChirpStack has no built-in PostgreSQL integration for application data, so this path needs a small consumer: an MQTT subscriber (or Node-RED style flow) that inserts each uplink into a TimescaleDB hypertable. The payoff is plain SQL in Grafana and easy joins against relational device metadata. The trade-off is the extra ingestion component you have to write and run.

ChirpStack, the time-series store and Grafana are three moving parts, plus the codecs and dashboards that make them useful. We build, wire and operate the full stack on European infrastructure as part of Grafana managed hosting alongside ChirpStack hosting, so you get dashboards instead of a backlog.