MikroTik Network Weathermap

An OVH/Cacti-style live network weathermap for MikroTik RouterOS devices. It polls your routers over the REST API, auto-discovers the topology from LLDP/MNDP neighbor data, computes live per-link utilization from interface counters, and serves an auto-refreshing HTML dashboard with color-coded links (green → yellow → red).

Features

• Physical-topology auto-discovery — links are matched physical port ↔ physical port by MAC, requiring bidirectional LLDP/MNDP agreement so flooded/indirect neighbors don't create phantom links (robust even with duplicate identities). Logical interfaces (bridges, VLANs/trunks) are ignored.
• Bond aggregation — member links (e.g. bonding1 over 2× SFP+) collapse into a single link with the summed bond speed (2×10G = 20G).
• Accurate capacities via /interface/ethernet/monitor negotiated rate (the configured speed is blank on auto-neg SFP+ ports).
• Live link colors & widths scaled by max(rx,tx) / capacity.
• D3 force-directed layout with zoom/pan, draggable/pinnable nodes, and a link TTL cache so transient neighbor-refresh gaps don't make nodes flicker.
• Hover tooltips with exact rx/tx, ports, capacity, and router model.
• --dump CLI to debug topology resolution.
• Multiple routers from a single YAML file. Pure stdlib HTTP server + requests/PyYAML. No database.

Quick start

pip install -r requirements.txt

# Preview with simulated data first (no router needed):
python3 weathermap.py --mock        # open http://127.0.0.1:8088

# Then wire up your real devices:
cp routers.example.yaml routers.yaml
$EDITOR routers.yaml
python3 weathermap.py                # http://127.0.0.1:8088

Bind to all interfaces / change port / poll faster:

python3 weathermap.py --host 0.0.0.0 --port 8088 --interval 3

Debug how links are being resolved (interfaces, neighbors, mutual matches, bonds, and the final link list):

python3 weathermap.py --dump --config routers.yaml

Docker

# Build & run with docker compose (mounts ./routers.yaml into the container)
docker compose up -d            # http://127.0.0.1:8088

# Or plain docker:
docker build -t mikrotik-weathermap .
docker run -d --name weathermap -p 8088:8088 \
  -v "$PWD/routers.yaml:/config/routers.yaml:ro" \
  mikrotik-weathermap

# Preview with simulated data, no config needed:
docker run --rm -p 8088:8088 mikrotik-weathermap --mock

Router setup (RouterOS v7+)

1. Enable the REST service (it rides on the www-ssl / www service):

   /ip/service/enable www-ssl       # https, recommended
   # or for plain http:
   /ip/service/enable www
   

2. Create a read-only user for the weathermap:

   /user/add name=weathermap group=read password=SECRET
   

3. Make sure neighbor discovery is on (it is by default) so topology can be auto-discovered:

   /ip/neighbor/discovery-settings/set discover-interface-list=all
   

Then add each device to routers.yaml. Use scheme: http / port: 80 if you only enabled the plain www service.

How utilization is computed

The server reads rx-byte / tx-byte from /interface on every poll and divides the byte delta by the elapsed time to get bits/sec. Capacity comes from the negotiated rate via /interface/ethernet/monitor, then the configured speed, then default_capacity_bps. Utilization = max(rx,tx) / cap. Bonded links use the summed bond speed as the denominator.

Files

File	Purpose
weathermap.py	Poller + HTTP/JSON server
index.html	Front-end (SVG weathermap, force layout)
routers.example.yaml	Copy to routers.yaml and edit
requirements.txt	Python deps

Notes

• verify_tls: false is the default because home routers use self-signed certs. Point it at a real cert and set true if you have one.
• Inter-router links require mutual MAC visibility; a port that only sees flooded (indirect) neighbors falls back to a stub rather than a false link.
• The front-end loads D3 from a CDN; the viewing machine needs internet access.
• To run headless without Docker, drop a systemd unit calling weathermap.py --host 0.0.0.0.