Unified dashboard for mixed Bitcoin SHA-256 ASIC fleets. One pane of glass for Bitaxe, BitForge, Avalon, and Nerd Miner devices — regardless of which firmware they run.
Most miner monitoring tools are single-protocol: they speak AxeOS REST and that's it. Fine if your whole fleet is Bitaxes, useless if you also have an Avalon Q on CGMiner TCP, plus Nerd Miners on a slightly different firmware, plus you want pool-side cross-reference for anything that escapes the IP list.
FleetSwarm runs three protocols in one poller:
| Hardware | Protocol | Port | What it reads |
|---|---|---|---|
| Bitaxe (Gamma, Supra, Ultra, Hex) | AxeOS REST | 80 | /api/system/info |
| BitForge | AxeOS REST | 80 | Same — AxeOS-derived firmware |
| Nerd Miner / NerdQAxe family | AxeOS REST | 80 | Same — AxeOS-compatible builds |
| Avalon Q (and other CGMiner ASICs) | CGMiner TCP | 4028 | estats + summary + pools |
| SoloStrike / Public Pool (optional) | HTTP API | varies | Worker cross-reference |
Subnet auto-discovery sweeps a /24 every cycle. New miner shows up in 30
seconds without any config changes.
- Fleet-wide hashrate, power, efficiency at the top of every page
- Configurable health thresholds for temperature, reject rate, and offline duration — the header pill flips green/orange/red based on rules you set
- Per-miner cards color-coded by class (BitForge cyan, Bitaxe amber, Avalon blue, Nerd magenta)
- Temperature alerting via color — green/orange/red against your bands
- Offline detection with diagonal-stripe styling + "last seen" timestamp
- 14-day SQLite history with per-miner time-series API endpoint
- Force poll button for instant refresh without waiting for the cycle
- Industrial terminal aesthetic — dense, glanceable, no decorative cruft
- Universal — no utility-specific or region-specific assumptions
- No telemetry. No phone-home. No cloud. Polls on your LAN, stores on your disk, serves to your browser. That's it.
-
Add the SoloStrike community app store: App Store → ⋯ → Community App Stores → Add and paste:
https://github.com/gbechtel-beck/umbrelsolostrike-app-store -
Find FleetSwarm under that store and click Install.
-
Open it. Go to Config, add your Avalon IP (if any), and tweak the alert thresholds if the defaults don't fit your hardware.
version: "3.7"
services:
fleetswarm:
image: ghcr.io/gbechtel-beck/fleetswarm:latest
restart: unless-stopped
ports:
- "8888:8888"
volumes:
- ./fleetswarm-data:/data
environment:
POLL_INTERVAL: "30"
DEFAULT_SUBNET: "192.168.1.0/24"
# Optional: set your timezone for accurate log timestamps
# TZ: America/New_Yorkdocker compose up -dOpen http://<host>:8888.
All settings are in the web UI under /config. Persisted to
/data/config.json so they survive container restarts and image updates.
| Field | Default | Notes |
|---|---|---|
| Subnet (CIDR) | 192.168.1.0/24 |
/24 only for the auto-scan |
| Subnet scan enabled | ✅ | Disable if you have a fixed IP list |
| Avalon IPs | empty | One per line — required for Avalon Q (CGMiner TCP) |
| Manual AxeOS IPs | empty | Useful for static IPs outside the scanned subnet |
| BTC payout address | empty | For Public Pool / SoloStrike worker cross-ref |
| Pool API base URL | empty | Same |
| Temp Warning (°C) | 65 | Card turns orange above this |
| Temp Critical (°C) | 75 | Card turns red, fleet alert flips critical |
| Reject Rate Warning (%) | 1.0 | Per-miner reject % that triggers warning |
| Offline Grace (min) | 5 | Offline-longer-than-this counts as critical |
Thresholds are universal — no hard-coded utility-specific or region-specific assumptions. Set them to whatever fits your hardware and tolerance.
| Endpoint | Method | Purpose |
|---|---|---|
/ |
GET | Dashboard UI |
/config |
GET | Settings UI |
/api/fleet |
GET | Current state for all miners + totals |
/api/miner/<id>/history?hours=24 |
GET | Per-miner time series |
/api/config |
GET / POST | Read or update config |
/api/poll-now |
POST | Force an immediate poll cycle |
/api/health |
GET | Heartbeat |
Every 30 seconds (configurable via POLL_INTERVAL), the background thread:
- Polls every IP in your explicit miner list via AxeOS HTTP
- Polls every IP in your Avalon list via CGMiner TCP
estats - Scans your subnet for any AxeOS responders not already known
- Writes one row per miner per cycle to the
samplestable - Marks any miner that didn't respond as offline
- Prunes samples older than 14 days
The web UI re-fetches /api/fleet every 5 seconds for snappy redraw without
adding load to the miners.
# Logs
docker logs -f umbrelsolostrike-fleetswarm_web_1 # on Umbrel
docker logs -f fleetswarm # plain docker
# Inspect the database
docker exec -it <container> sqlite3 /data/fleet.db
sqlite> SELECT hostname, kind, online FROM miners;
sqlite> SELECT hostname, hashrate_ghs, ts FROM samples
...> JOIN miners ON miners.id = samples.miner_id
...> ORDER BY ts DESC LIMIT 20;
# Force a poll cycle
curl -X POST http://<host>:8888/api/poll-now
# Reset everything (config + DB)
docker stop <container>
rm /path/to/data/fleet.db /path/to/data/config.json
docker start <container>- Per-miner detail page with hashrate sparkline
- Pool websocket cross-reference (Public Pool / SoloStrike)
- Discord webhook alerts on offline > N minutes
- Auto-detect Avalon during subnet scan (probe TCP 4028)
- CSV export for history / accounting
- Optional cost tracking module — flat-rate or time-of-use, fully user-defined (off by default, no region assumptions)
- Per-pool grouping in addition to per-class
git clone https://github.com/gbechtel-beck/fleetswarm
cd fleetswarm
docker build -t fleetswarm:dev .
docker run -d -p 8888:8888 -v $(pwd)/data:/data fleetswarm:devTo trigger a release build that publishes to Docker Hub:
git tag v0.1.0
git push origin v0.1.0The GitHub Actions workflow will build a multi-arch image (amd64 + arm64) and
push it to Docker Hub. The workflow's last step prints the digest you need to
pin in the Umbrel app's docker-compose.yml.
MIT. Do whatever you want with it. Attribution appreciated, not required.
- The Bitaxe community for AxeOS and the open-source mining renaissance
- Skot for kicking the whole thing off
- Stu1958 for the ESP32-CYD miner-swarm-dashboard firmware that inspired the name and the spirit of "look at it and instantly know your fleet is healthy." Different architecture, same mission.
- Everyone in the OSMU Discord