Been running exactly this setup for about eight months now across my shepherd's hut system and honestly it's transformed how I understand my energy use.
I've got a Raspberry Pi 4 pulling data from a Victron Cerbo GX via the local MQTT broker — no cloud dependency, everything stays on-site. Paired it with InfluxDB and Grafana, and now I've got beautiful dashboards showing SOC, solar yield, battery voltage trends, and load consumption going back months.
The real win came when I started charging my van from the hut's system. Being able to overlay the EV charging load against solar production meant I could fine-tune exactly when to start a charge session to avoid hammering the Fogstar batteries during cloudy patches.
A few things worth knowing if you're setting this up:
- VRM Local API works brilliantly without needing internet
- The dbus-mqtt bridge on the Cerbo is your friend — documentation is a bit buried but it's solid once you find it
- Set your Pi to write to an external SSD rather than the SD card — learned that one the hard way after a corruption event
The only faff was getting the Grafana alerts tuned properly so I wasn't getting pinged at midnight about perfectly normal overnight discharge cycles.
Anyone else gone down this route? Curious whether others are pulling in weather forecast data to anticipate low-yield days — I've been eyeing up the Open-Meteo API for this but haven't pulled the trigger yet. Also interested if anyone's managed to get consumption data from a Renogy charge controller alongside Victron kit in the same dashboard.