2022-01-15 12:41:31 -05:00
# Reverse Proxy Setup Guide
It is recommended to run `nostr-rs-relay` behind a reverse proxy such
2023-02-24 11:06:30 -05:00
as `haproxy` , `nginx` or `traefik` to provide TLS termination. Simple examples
for `haproxy` , `nginx` and `traefik` configurations are documented here.
2022-01-15 12:41:31 -05:00
## Minimal HAProxy Configuration
Assumptions:
* HAProxy version is `2.4.10` or greater (older versions not tested).
* Hostname for the relay is `relay.example.com` .
* Your relay should be available over wss://relay.example.com
* Your (NIP-11) relay info page should be available on https://relay.example.com
* SSL certificate is located in `/etc/certs/example.com.pem` .
* Relay is running on port 8080.
* Limit connections to 400 concurrent.
* HSTS (HTTP Strict Transport Security) is desired.
* Only TLS 1.2 or greater is allowed.
```
global
ssl-default-bind-ciphersuites TLS_AES_128_GCM_SHA256:TLS_AES_256_GCM_SHA384:TLS_CHACHA20_POLY1305_SHA256
ssl-default-bind-options prefer-client-ciphers no-sslv3 no-tlsv10 no-tlsv11 no-tls-tickets
frontend fe_prod
mode http
bind :443 ssl crt /etc/certs/example.com.pem alpn h2,http/1.1
bind :80
http-request set-header X-Forwarded-Proto https if { ssl_fc }
redirect scheme https code 301 if !{ ssl_fc }
Fork Sync: Update from parent repository (#3)
* improvement: use appropriate paths for systemd example
* improvement: add a configurable postgres write conn string
This adds a new configurable connection string for postgres writes.
* improvement: document pg connection_write config
* build: upgrade checkout action for github ci
* perf: use standard allocator, limit sqlite mmap to 4GB
This is an experimental change to see if we can reduce memory usage
with large SQLite databases. If successful, we'll do this again and
further reduce the database mmap size.
This will cause greater use of the page cache, but that is more easily
reclaimed by the kernel, and should reduce memory pressure, as well as
making it clearer how much memory the application is actually using
for connections, subscriptions, etc.
* docs: reformatting
* docs: allow host header prefix matching, required for Damus compatibility
* perf: disable sqlite mmap to reduce memory pressure
* perf: switch to jemalloc allocator
* docs: helpful ubuntu packages for building
* perf: reduce SQLite connection count and idle lifetime
On lightly loaded relays, we free up memory faster by letting idle
connections be reclaimed in 10 seconds instead of the default 10
minutes. This also sets the minimum to zero connections, instead of
always trying to hold one open.
---------
Co-authored-by: Petr Kracik <petrkr@petrkr.net>
Co-authored-by: Kieran <kieran@harkin.me>
Co-authored-by: Greg Heartsfield <scsibug@imap.cc>
2023-05-10 04:30:18 -04:00
acl host_relay hdr(host) -i -m beg relay.example.com
2022-01-15 12:41:31 -05:00
use_backend relay if host_relay
# HSTS (1 year)
http-response set-header Strict-Transport-Security max-age=31536000
backend relay
mode http
2022-01-19 22:19:12 -05:00
timeout connect 5s
timeout client 50s
timeout server 50s
timeout tunnel 1h
timeout client-fin 30s
2022-01-15 12:41:31 -05:00
option tcp-check
default-server maxconn 400 check inter 20s fastinter 1s
2022-01-15 12:43:12 -05:00
server relay 127.0.0.1:8080
2022-01-15 12:41:31 -05:00
```
2022-08-20 12:31:49 -04:00
### HAProxy Notes
2022-01-15 12:41:31 -05:00
You may experience WebSocket connection problems with Firefox if
HTTP/2 is enabled, for older versions of HAProxy (2.3.x). Either
disable HTTP/2 (`h2`), or upgrade HAProxy.
2022-08-20 12:31:49 -04:00
## Bare-bones Nginx Configuration
Assumptions:
* `Nginx` version is `1.18.0` (other versions not tested).
* Hostname for the relay is `relay.example.com` .
* SSL certificate and key are located at `/etc/letsencrypt/live/relay.example.com/` .
* Relay is running on port `8080` .
```
http {
server {
listen 443 ssl;
server_name relay.example.com;
ssl_certificate /etc/letsencrypt/live/relay.example.com/fullchain.pem;
ssl_certificate_key /etc/letsencrypt/live/relay.example.com/privkey.pem;
2023-01-02 20:08:06 -05:00
ssl_protocols TLSv1.3 TLSv1.2;
ssl_prefer_server_ciphers on;
ssl_ecdh_curve secp521r1:secp384r1;
ssl_ciphers EECDH+AESGCM:EECDH+AES256;
# Optional Diffie-Helmann parameters
# Generate with openssl dhparam -out /etc/ssl/certs/dhparam.pem 4096
#ssl_dhparam /etc/ssl/certs/dhparam.pem;
ssl_session_cache shared:TLS:2m;
ssl_buffer_size 4k;
# OCSP stapling
ssl_stapling on;
ssl_stapling_verify on;
resolver 1.1.1.1 1.0.0.1 [2606:4700:4700::1111] [2606:4700:4700::1001]; # Cloudflare
# Set HSTS to 365 days
add_header Strict-Transport-Security 'max-age=31536000; includeSubDomains; preload' always;
2022-08-20 12:31:49 -04:00
keepalive_timeout 70;
location / {
proxy_pass http://localhost:8080;
proxy_http_version 1.1;
2023-02-05 18:17:13 -05:00
proxy_read_timeout 1d;
proxy_send_timeout 1d;
2022-08-20 12:31:49 -04:00
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection "Upgrade";
proxy_set_header Host $host;
}
}
}
```
### Nginx Notes
2023-02-15 03:28:42 -05:00
The above configuration was tested on `nginx` `1.18.0` on `Ubuntu` `20.04` and `22.04`
2022-08-20 12:31:49 -04:00
For help installing `nginx` on `Ubuntu` , see [this guide ](https://www.digitalocean.com/community/tutorials/how-to-install-nginx-on-ubuntu-20-04 ).
For guidance on using `letsencrypt` to obtain a cert on `Ubuntu` , including an `nginx` plugin, see [this post ](https://www.digitalocean.com/community/tutorials/how-to-secure-nginx-with-let-s-encrypt-on-ubuntu-20-04 ).
2023-02-24 11:06:30 -05:00
## Example Traefik Configuration
Assumptions:
* `Traefik` version is `2.9` (other versions not tested).
* `Traefik` is used for provisioning of Let's Encrypt certificates.
* `Traefik` is running in `Docker` , using `docker compose` and labels for the static configuration. An equivalent setup useing a Traefik config file is possible too (but not covered here).
* Strict Transport Security is enabled.
* Hostname for the relay is `relay.example.com` , email adres for ACME certificates provider is `name@example.com` .
Fork Sync: Update from parent repository (#3)
* improvement: use appropriate paths for systemd example
* improvement: add a configurable postgres write conn string
This adds a new configurable connection string for postgres writes.
* improvement: document pg connection_write config
* build: upgrade checkout action for github ci
* perf: use standard allocator, limit sqlite mmap to 4GB
This is an experimental change to see if we can reduce memory usage
with large SQLite databases. If successful, we'll do this again and
further reduce the database mmap size.
This will cause greater use of the page cache, but that is more easily
reclaimed by the kernel, and should reduce memory pressure, as well as
making it clearer how much memory the application is actually using
for connections, subscriptions, etc.
* docs: reformatting
* docs: allow host header prefix matching, required for Damus compatibility
* perf: disable sqlite mmap to reduce memory pressure
* perf: switch to jemalloc allocator
* docs: helpful ubuntu packages for building
* perf: reduce SQLite connection count and idle lifetime
On lightly loaded relays, we free up memory faster by letting idle
connections be reclaimed in 10 seconds instead of the default 10
minutes. This also sets the minimum to zero connections, instead of
always trying to hold one open.
---------
Co-authored-by: Petr Kracik <petrkr@petrkr.net>
Co-authored-by: Kieran <kieran@harkin.me>
Co-authored-by: Greg Heartsfield <scsibug@imap.cc>
2023-05-10 04:30:18 -04:00
* ipv6 is enabled, a viable private ipv6 subnet is specified in the example below.
2023-02-24 11:06:30 -05:00
* Relay is running on port `8080` .
```
version: '3'
networks:
nostr:
enable_ipv6: true
ipam:
config:
- subnet: fd00:db8:a::/64
gateway: fd00:db8:a::1
services:
traefik:
image: traefik:v2.9
networks:
nostr:
command:
- "--log.level=ERROR"
# letsencrypt configuration
- "--certificatesResolvers.http.acme.email==name@example.com"
- "--certificatesResolvers.http.acme.storage=/certs/acme.json"
- "--certificatesResolvers.http.acme.httpChallenge.entryPoint=http"
# define entrypoints
- "--entryPoints.http.address=:80"
- "--entryPoints.http.http.redirections.entryPoint.to=https"
- "--entryPoints.http.http.redirections.entryPoint.scheme=https"
- "--entryPoints.https.address=:443"
- "--entryPoints.https.forwardedHeaders.insecure=true"
- "--entryPoints.https.proxyProtocol.insecure=true"
# docker provider (get configuration from container labels)
- "--providers.docker.endpoint=unix:///var/run/docker.sock"
- "--providers.docker.exposedByDefault=false"
- "--providers.file.directory=/config"
- "--providers.file.watch=true"
ports:
- "80:80"
- "443:443"
volumes:
- "/var/run/docker.sock:/var/run/docker.sock:ro"
- "$(pwd)/traefik/certs:/certs"
- "$(pwd)/traefik/config:/config"
logging:
driver: "local"
restart: always
# example nostr config. only labels: section is relevant for Traefik config
nostr:
image: nostr-rs-relay:latest
container_name: nostr-relay
networks:
nostr:
restart: always
user: 100:100
volumes:
- '$(pwd)/nostr/data:/usr/src/app/db:Z'
- '$(pwd)/nostr/config/config.toml:/usr/src/app/config.toml:ro,Z'
labels:
- "traefik.enable=true"
- "traefik.http.routers.nostr.entrypoints=https"
- "traefik.http.routers.nostr.rule=Host(`relay.example.com`)"
- "traefik.http.routers.nostr.tls.certresolver=http"
- "traefik.http.routers.nostr.service=nostr"
- "traefik.http.services.nostr.loadbalancer.server.port=8080"
- "traefik.http.services.nostr.loadbalancer.passHostHeader=true"
- "traefik.http.middlewares.nostr.headers.sslredirect=true"
- "traefik.http.middlewares.nostr.headers.stsincludesubdomains=true"
- "traefik.http.middlewares.nostr.headers.stspreload=true"
- "traefik.http.middlewares.nostr.headers.stsseconds=63072000"
- "traefik.http.routers.nostr.middlewares=nostr"
```
### Traefik Notes
Fork Sync: Update from parent repository (#3)
* improvement: use appropriate paths for systemd example
* improvement: add a configurable postgres write conn string
This adds a new configurable connection string for postgres writes.
* improvement: document pg connection_write config
* build: upgrade checkout action for github ci
* perf: use standard allocator, limit sqlite mmap to 4GB
This is an experimental change to see if we can reduce memory usage
with large SQLite databases. If successful, we'll do this again and
further reduce the database mmap size.
This will cause greater use of the page cache, but that is more easily
reclaimed by the kernel, and should reduce memory pressure, as well as
making it clearer how much memory the application is actually using
for connections, subscriptions, etc.
* docs: reformatting
* docs: allow host header prefix matching, required for Damus compatibility
* perf: disable sqlite mmap to reduce memory pressure
* perf: switch to jemalloc allocator
* docs: helpful ubuntu packages for building
* perf: reduce SQLite connection count and idle lifetime
On lightly loaded relays, we free up memory faster by letting idle
connections be reclaimed in 10 seconds instead of the default 10
minutes. This also sets the minimum to zero connections, instead of
always trying to hold one open.
---------
Co-authored-by: Petr Kracik <petrkr@petrkr.net>
Co-authored-by: Kieran <kieran@harkin.me>
Co-authored-by: Greg Heartsfield <scsibug@imap.cc>
2023-05-10 04:30:18 -04:00
Traefik will take care of the provisioning and renewal of certificates. In case of an ipv4-only relay, simply detele the `enable_ipv6:` and `ipam:` entries in the `networks:` section of the docker-compose file.
