The sozu executable can be used to start the proxy, and configure it: adding new backend servers, reading metrics, etc. It talks to the currently running proxy through a unix socket.
You can specify its path by adding to your config.toml
:
command_socket = "path/to/your/command_folder/sock"
First you need to create a new cluster with an id and a load balancing policy (roundrobin or random):
sozu --config /etc/sozu/config.toml cluster add --id <my_cluster_id> --load-balancing-policy roundrobin
It won't show anything but you can verify that the cluster has been added successfully by querying sozu:
sozu --config /etc/sozu/config.toml query clusters
Then you need to add a backend:
sozu --config /etc/sozu/config.toml backend add --address 127.0.0.1:3000 --backend-id <my_backend_id> --id <my_cluster_id>
And an http listener:
sozu --config /etc/sozu/config.toml listener http add --address 0.0.0.0:80 --tls-versions TLSv1.2 --tls-cipher-list ECDHE-ECDSA-AES256-GCM-SHA384 --tls-cipher-suites TLS_AES_256_GCM_SHA384 --tls-signature-algorithms ECDSA+SHA512 --tls-groups-list x25519 --expect-proxy
Finally you have to create a frontend to allow sozu to send traffic from the listener to your backend:
sozu --config /etc/sozu/config.toml frontend http add --address 0.0.0.0:80 --hostname <my_cluster_hostname> id <my_cluster_id>
And an https listener:
sozu --config /etc/sozu/config.toml listener https add --address 0.0.0.0:443
Finally you have to create a frontend to allow sozu to send traffic from the listener to your backend:
sozu --config /etc/sozu/config.toml frontend https add --address 0.0.0.0:443 --hostname <my_cluster_hostname> id <my_cluster_id>
It shows a list of workers and show information about their statuses.
sozu --config /etc/sozu/config.toml status
It will show global statistics about sozu, workers and clusters metrics.
sozu --config /etc/sozu/config.toml query metrics
If sozu configurations (clusters, frontends & backends) are not written in the config file, you can save sozu state to restore it later.
sozu --config /etc/sozu/config.toml state save --file state.json
Then shutdown gracefully sozu:
sozu --config /etc/sozu/config.toml shutdown
Restart sozu and restore its state:
sozu --config /etc/sozu/config.toml state load --file state.json
You should be able to request your cluster like before the shutdown.
This CLI command:
sozu --config /path/to/config.toml events
listens to events sent by Sōzu workers whenever a backend is down, up again, or when no backend is available.