Monitoring energy usage with smart plugs, Prometheus and Grafana

Original Article ↗ Hacker News Discussion ↗

Home energy monitoring is increasingly being built from off‑the‑shelf smart plugs, whole‑home current transformers and smart meters, with data funneled into Prometheus and visualized in Grafana or alternatives like Home Assistant. Commenters weigh trade‑offs between Wi‑Fi vs Zigbee/Z‑Wave devices, cloud‑tied products vs fully local control, and turnkey platforms vs DIY stacks, stressing reliability, safety (especially of cheap mains‑rated gear), and ease of integration. There is also interest in going beyond charts to automation and SCADA‑style control, but opinions differ on whether that belongs in Grafana, Home Assistant, or more programmable tools like Node‑RED.

Home Assistant vs. Custom Monitoring Stacks

  • Many argue that Home Assistant (HA) already provides energy monitoring, graphs, and automations “out of the box,” making Prometheus+Grafana overkill for basic home use.
  • Others report bad experiences with HAOS / supervised Docker setups (laggy UI, failures that are hard to debug inside containers) and prefer traditional package-managed installs or NixOS-style deployments.
  • There’s tension around HA’s distribution philosophy: some see containers as a reproducible “stable state,” others dislike being locked into a bespoke mini‑distro and perceive hostility toward third‑party packaging (e.g., NixOS).
  • Several note that HA Core / Container now exist as more flexible options, but mistrust remains for some users who prefer “MQTT-first” architectures and separate logging/automation stacks.

Grafana, Automation, and SCADA Direction

  • Grafana is praised for deeper analysis, aggregations, and comparisons versus HA’s “at a glance” graphs, especially by off‑grid users.
  • Some run both: HA for device integration and automations; Prometheus/Influx + Grafana for serious visualization.
  • A Grafana developer mentions ongoing work to push Grafana toward IoT/SCADA control via “data sinks” (e.g., HTTP, MQTT, Modbus) and interactive Canvas panels; others suggest looking at Node‑RED’s ecosystem as inspiration.

Protocols and Network Choices

  • Several recommend Zigbee or Z‑Wave plugs over WiFi for scale and reliability; others report 20+ WiFi/ESP devices working fine, especially when 2.4 GHz is reserved mostly for IoT.
  • Debate exists over whether WiFi unreliability is due to cheap devices, poor RF environments, or infrastructure issues.
  • Powerline networking is criticized for RF noise; some reminisce about older powerline/X10 control being simple and robust.

Hardware Choices, Calibration, and Safety

  • Many plug recommendations: Shelly, Athom (ESPHome/Tasmota), TP‑Link/Tapo, Innr, Robb, ThirdReality, IKEA, Gosund, various Zigbee outlets, etc.; experiences vary, including premature failures on some cheap brands.
  • Calibration of Tasmota/ESPHome power monitoring is raised; others rely on vendors claiming “pre‑calibrated.”
  • Strong warnings against cheap AliExpress main‑breakers and relays being used at full rated current; CT clamps and dedicated monitors (Emporia Vue, IoTaWatt, Shelly 3EM) are seen as safer for whole‑home or high‑amp circuits.
  • Overvoltage behavior differs by brand and may conflict with local grid conditions, requiring automations to recover tripped devices.

Whole‑Home and Utility‑Based Monitoring

  • Approaches include: multi‑CT panel monitors (Emporia Vue, IoTaWatt), Zigbee breakers, smart‑meter interfaces (P1/HAN/DLMS readers, vendor dongles, Eagle gateways), and DIY ESP32 sensors on meter LEDs/disks.
  • Some countries’ utilities provide hourly or near‑hourly consumption data via portals; others only offer coarse, delayed summaries, driving interest in local monitoring.