TeamMapper supports deployment behind a reverse proxy. Here's an example configuration using nginx with docker compose, though other reverse proxy setups will work as well. Add these lines to your docker-compose.yml:
nginx:
image: nginx:latest
ports:
- 80:80
volumes:
- ./nginx/nginx.conf:/etc/nginx/nginx.conf:ro
Then create a nginx.conf file in the folder nginx, with one of the following setups.
server {
listen 80;
listen [::]:80;
server_name teammapper.lan.DOMAIN.tld;
return 301 https://$host$request_uri;
}
server {
listen 443 ssl;
listen [::]:443 ssl;
server_name teammapper.lan.DOMAIN.tld;
# certificates for ssl, change according to your setup
# ssl_certificate /etc/letsencrypt/live/lan.DOMAIN.tld/fullchain.pem;
# ssl_certificate_key /etc/letsencrypt/live/lan.DOMAIN.tld/privkey.pem;
location / {
proxy_pass http://teammapper:3000/;
proxy_http_version 1.1;
proxy_set_header Host $http_host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header Access-Control-Allow-Origin *;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection "Upgrade";
}
}
When hosting TeamMapper under a path (e.g. /teammapper/), the <base href> tag in the HTML needs to be updated to match that path. This ensures TeamMapper can correctly resolve relative URLs for assets, API calls, and navigation. For nginx, this can be achieved using the sub_filter directive to dynamically replace the base href. For Apache, the equivalent functionality is provided by the mod_substitute module.
events {
worker_connections 1024;
}
http {
server {
listen 80;
server_name localhost;
location /teammapper/ {
proxy_pass http://teammapper:3000/;
sub_filter '<base href="/"/>' '<base href="/teammapper/" />';
sub_filter_types text/html;
sub_filter_once on;
# Add proper proxy headers
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
# WebSocket support
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection "upgrade";
}
}
}
These limits apply independently of FEATURE_YJS_RATE_LIMITING.
| Backend variable | Default | Unit | Limit |
|---|---|---|---|
YJS_MAP_MAX_BYTES |
5242880 | bytes | Map state, including history |
YJS_MAP_MAX_ENTRIES |
100000 | Yjs entries | Entries in the map |
YJS_MAP_MAX_NODES |
10000 | nodes | Nodes in the map |
YJS_MESSAGE_WINDOW_MS |
10000 | milliseconds | Rate window and reset cooldown |
YJS_MESSAGE_MAX_COUNT |
100 | messages | Per window, per connection and per map |
YJS_MESSAGE_MAX_BYTES |
5242880 | bytes | Per window, per connection and per map |
When a limit is reached:
- Map size (
YJS_MAP_*), or invalid content: the backend discards the live map and disconnects every peer. Peers reload the saved map after the cooldown, and unsaved edits may be lost. - Messages per connection, all messages including presence: the backend disconnects that connection.
- Messages per map, edits only: the backend disconnects the connection that sent the most in the window, and the other peers keep working. At twice the limit it disconnects every sender.
A disconnected client reconnects after a few hundred milliseconds and resends any edits the backend did not apply.