Skip to main content

Installing Zilla

Installing Zilla

Running Zilla via Docker

Run the latest Zilla release with the default empty configuration via docker.

docker run start -v

The output should display the zilla config and started to know zilla is ready for traffic.

// default Zilla config
name: default

// Zilla status

Specify your own zilla.yaml file.

Docker 23
docker run -v ./zilla.yaml:/etc/zilla/zilla.yaml start -v

Running Zilla via Helm

Go to the Zilla artifacthubopen in new window page to find out more on how to install Zilla using Helm


helm install zilla oci:// --namespace zilla --create-namespace --wait \
    --values values.yaml \
    --set-file zilla\\.yaml=zilla.yaml


Zilla specific configuration is in the zilla.yaml file which can be included in the helm install by adding --set-file zilla\\.yaml=zilla.yaml to your command.

zilla.yaml Auto Reconfigure

Zilla loads the configuration from the zilla.yaml file on startup and logs the configured settings. Restarting Zilla or its container may not be an option, so Zilla creates a file watcher to detect changes to the file and reloads the config if a change is detected.

Errors and misconfigured parts of the zilla.yaml file are detected by zilla and reported via stdout. The original config remains in place and can only be replaced by a valid config.

This feature is demonstrated in the above Helm install command. Running a helm update ... with changes to the zilla.yaml, k8s will update the config map, which writes the new content into the running pods. Zilla will detect those file changes and load the new config.

Auto Scaling

Zilla will start workers that default to the CPU cores it is allowed to use. This makes horizontal scaling easy with a 1:1 ratio of instances to workers. Any of the default scaling metrics based on server CPU usage will enable Zilla to handle traffic spikes. Additionally, Zilla Telemetry configuration provides more data when determining how to scale. The Prometheus autoscale exampleopen in new window demonstrates using metrics from the Prometheus exporter to horizontally scale Zilla on k8s.