Integrations

ESPGeiger exposes radiation data through several channels (MQTT, Webhook, JSON endpoint). This page shows how to connect these to common monitoring and automation platforms.

JSON Endpoint

The built-in web server exposes a JSON endpoint at /json:

http://192.168.1.100/json

Response:

{"ut":5025,"c":42.00,"s":0.28,"c5":41.50,"c15":40.90,"cs":0.70,"r":151.0,"tc":9876,"mem":19088,"rssi":-45}

Field	Description
ut	Uptime in seconds
c	Current CPM
s	Current μSv/h
c5	5-minute smoothed CPM
c15	15-minute smoothed CPM
cs	Current CPS
r	μSv conversion ratio
tc	Total clicks since boot
mem	Free heap memory in bytes
rssi	WiFi signal strength in dBm
hv	HV reading in volts (ESPGeiger-HW only)
loss	Packet loss on the feed, as a percentage (UDP-receiver builds only; appears once a producer has been heard)

Raw / diagnostic values (?raw=1)

mem is lightly smoothed so a single read can’t catch a transient dip and report a phantom drop, and the device-load fields (tick, t_max, lps) are left off the default response to keep the frequently-polled /json small. Append ?raw=1 for diagnostics:

http://192.168.1.100/json?raw=1

Raw mode reports mem un-smoothed and adds:

Field	Description
tick	Smoothed microseconds spent in the once-per-second housekeeping pass. Lower is better.
t_max	Worst-case tick in the last 60 seconds.
lps	Main-loop iterations per second, un-smoothed (last full second).
frag	Heap fragmentation, percent
lfb	Largest contiguous free block in bytes
lfblow	Largest-free-block low-water mark since boot, in bytes

lfb / lfblow are the honest measure of memory health: free heap can look fine while fragmentation quietly shrinks the largest block you can actually allocate.

GeigerLog

GeigerLog is a free cross-platform log/plot tool for Geiger counters. ESPGeiger exposes a GeigerLog-compatible endpoint at /lastdata that works with GeigerLog’s built-in WiFiServer device (the one that polls a device URL).

Configure GeigerLog

1. In GeigerLog open Device → WiFiServer → Config WiFiServer Device.
2. Set the IP address and port to the ESPGeiger device (port 80).
3. Point the data URL at /lastdata.
4. Save, then connect the WiFiServer device.

The /lastdata endpoint returns a single comma-separated line in GeigerLog’s 12-variable order. Fields the device can’t provide are left empty:

22.50, 0.37, 21.90, , 21.40, , , , 24.50, 1013.00, 48.00, 0.0720

Position	GeigerLog variable	ESPGeiger value
1	CPM	Current CPM
2	CPS	Current CPS
3	CPM1st	5-minute CPM
5	CPM2nd	15-minute CPM
9	Temp	Temperature, if an environment sensor is fitted
10	Press	Pressure, if fitted
11	Humid	Humidity, if fitted
12	Xtra	HV in volts (ESPGeiger-HW), otherwise dose rate in µSv/h

Positions 4, 6, 7 and 8 (CPS1st / CPS2nd / CPM3rd / CPS3rd) are always empty.

Notes

• Polling is driven by GeigerLog - ESPGeiger responds on each request, so cadence is whatever you configure on the GeigerLog side.
• The endpoint is always available; there is no toggle in the ESPGeiger portal.
• Environment fields appear only when a sensor is detected; otherwise those positions stay empty.

Prometheus (native /metrics)

ESPGeiger exposes a native Prometheus text exposition endpoint at /metrics. No proxy or exporter needed.

Scrape config

# prometheus.yml
scrape_configs:
  - job_name: espgeiger
    metrics_path: /metrics
    scrape_interval: 30s
    static_configs:
      - targets:
          - 192.168.1.100   # your device(s)
          - 192.168.1.101

For automatic discovery, each device advertises a _geiger._tcp mDNS service on port 80 (alongside the standard _http._tcp), so you can point your service-discovery layer at _geiger._tcp to find every ESPGeiger on the network.

Sample output

# HELP device_info Device metadata; value always 1. Join via on(chipid) for naming.
# TYPE device_info gauge
device_info{chipid="d5536d",name="Garage",model="ESPGeiger-HW",ver="0.11.0",env="espgeigerhw"} 1

# HELP geiger_cpm Counts per minute (window label selects averaging period)
# TYPE geiger_cpm gauge
geiger_cpm{chipid="d5536d",window="1m"} 31.50
geiger_cpm{chipid="d5536d",window="5m"} 30.20
geiger_cpm{chipid="d5536d",window="15m"} 29.80

# HELP geiger_cps Instantaneous counts per second
geiger_cps{chipid="d5536d"} 0.52

# HELP geiger_usv_per_hour Dose rate in microsieverts per hour
geiger_usv_per_hour{chipid="d5536d"} 0.2086

# HELP geiger_lifetime_clicks_total Total clicks counted over device life
# TYPE geiger_lifetime_clicks_total counter
geiger_lifetime_clicks_total{chipid="d5536d"} 1845231

# HELP device_uptime_seconds Seconds since boot
device_uptime_seconds{chipid="d5536d"} 90523

# HELP device_free_heap_bytes Free heap memory
device_free_heap_bytes{chipid="d5536d"} 19432

# HELP device_wifi_rssi_dbm WiFi signal strength
device_wifi_rssi_dbm{chipid="d5536d"} -54

# HELP geiger_pulse_interval_seconds Inter-pulse intervals in seconds, log2 buckets
# TYPE geiger_pulse_interval_seconds histogram
geiger_pulse_interval_seconds_bucket{chipid="d5536d",le="0.000064"} 0
geiger_pulse_interval_seconds_bucket{chipid="d5536d",le="0.000128"} 0
...
geiger_pulse_interval_seconds_bucket{chipid="d5536d",le="+Inf"} 1845231
geiger_pulse_interval_seconds_count{chipid="d5536d"} 1845231

Label model

Operational metrics carry only chipid - a stable device ID that never changes. Friendly name, model and firmware version live in the separate device_info metric. This means:

• Renaming a device in the portal does not fork its historical time series.
• Dashboard joins use on(chipid) group_left(name) to bring in the friendly name.

PromQL examples

# Current CPM with friendly name as legend
geiger_cpm{window="1m"} * on(chipid) group_left(name) device_info

# Dose rate averaged over 1 hour
avg_over_time(geiger_usv_per_hour[1h])
  * on(chipid) group_left(name) device_info

# Alert when any device exceeds 1 µSv/h sustained for 5 min
- alert: RadiationHigh
  expr: avg_over_time(geiger_usv_per_hour[5m]) > 1
  labels:
    severity: warning
  annotations:
    summary: " dose rate sustained above 1 µSv/h"

# Devices that haven't reported in 5 min
time() - timestamp(geiger_cpm{window="1m"}) > 300

Notes

• The endpoint is always available; there is no toggle in the portal.
• Histogram bucket boundaries are powers of two starting at 64 µs (inter-pulse interval). The top bucket is +Inf.
• geiger_lifetime_clicks_total is a Prometheus counter (monotonically increasing); use rate() for click rate.

---

Prometheus via Telegraf

If you’d rather pull from the /json endpoint (eg you already run Telegraf, or you want to add fields not in /metrics), Telegraf can poll JSON and expose Prometheus metrics:

# telegraf.conf

[[inputs.http]]
  urls = ["http://192.168.1.100/json"]
  method = "GET"
  timeout = "5s"
  data_format = "json_v2"
  name_override = "espgeiger"
  [[inputs.http.tags]]
    device = "espgeiger-xxxxxx"
  [[inputs.http.json_v2]]
    [[inputs.http.json_v2.field]]
      path = "c"
      rename = "cpm"
      type = "float"
    [[inputs.http.json_v2.field]]
      path = "s"
      rename = "usv_h"
      type = "float"
    [[inputs.http.json_v2.field]]
      path = "c5"
      rename = "cpm5"
      type = "float"
    [[inputs.http.json_v2.field]]
      path = "c15"
      rename = "cpm15"
      type = "float"
    [[inputs.http.json_v2.field]]
      path = "cs"
      rename = "cps"
      type = "float"
    [[inputs.http.json_v2.field]]
      path = "tc"
      rename = "total_clicks"
      type = "int"
    [[inputs.http.json_v2.field]]
      path = "ut"
      rename = "uptime_seconds"
      type = "int"
    [[inputs.http.json_v2.field]]
      path = "mem"
      rename = "free_heap"
      type = "int"
    [[inputs.http.json_v2.field]]
      path = "rssi"
      rename = "wifi_rssi"
      type = "int"

[[outputs.prometheus_client]]
  listen = ":9273"

Prometheus can then scrape http://<telegraf-host>:9273/metrics.

Prometheus via MQTT Exporter

If you are already publishing via MQTT, the mqtt2prometheus exporter converts MQTT topics into Prometheus metrics:

# mqtt2prometheus.yaml
mqtt:
  server: tcp://mqtt-broker:1883
  topic_path: ESPGeiger-+/stat/+
  device_id_regex: "ESPGeiger-(?P<deviceid>[0-9a-f]+)/stat/.+"
  qos: 0
metrics:
  - prom_name: espgeiger_cpm
    mqtt_name: CPM
    help: Current counts per minute
    type: gauge
  - prom_name: espgeiger_usv
    mqtt_name: uSv
    help: Current microsieverts per hour
    type: gauge
  - prom_name: espgeiger_cpm5
    mqtt_name: CPM5
    help: 5-minute smoothed CPM
    type: gauge

InfluxDB via Telegraf

To write directly to InfluxDB instead of Prometheus:

[[inputs.http]]
  urls = ["http://192.168.1.100/json"]
  data_format = "json_v2"
  name_override = "espgeiger"
  # ... field config as above ...

[[outputs.influxdb_v2]]
  urls = ["http://influxdb:8086"]
  token = "YOUR_TOKEN"
  organization = "your-org"
  bucket = "radiation"

Node-RED

Node-RED can poll the JSON endpoint and route data anywhere:

[
  {
    "id": "poll",
    "type": "inject",
    "repeat": "60",
    "topic": "",
    "payload": "",
    "payloadType": "date"
  },
  {
    "id": "http",
    "type": "http request",
    "method": "GET",
    "url": "http://192.168.1.100/json",
    "ret": "obj"
  }
]

Wire poll → http → your destination (MQTT, database, dashboard, webhook, etc.).

Home Assistant

If using MQTT with Home Assistant autodiscovery enabled, ESPGeiger sensors appear automatically. See Home Assistant for details.

For REST polling without MQTT:

# configuration.yaml
rest:
  - resource: http://192.168.1.100/json
    scan_interval: 30
    sensor:
      - name: "Radiation CPM"
        value_template: ""
        unit_of_measurement: "CPM"
      - name: "Radiation uSv/h"
        value_template: ""
        unit_of_measurement: "μSv/h"
        device_class: "radiation"
      - name: "ESPGeiger WiFi RSSI"
        value_template: ""
        unit_of_measurement: "dBm"
        device_class: "signal_strength"
      - name: "ESPGeiger Free Heap"
        value_template: ""
        unit_of_measurement: "B"

Grafana Dashboard

Once data is in Prometheus or InfluxDB, Grafana can visualise it. A simple dashboard query for Prometheus:

# Current CPM
espgeiger_cpm

# 5 minute average
avg_over_time(espgeiger_cpm[5m])

# Alert when above threshold
espgeiger_cpm > 100