Dynamic Config file generation can include creation of

1. Config file structure:

   The config files for e.g redis.yaml might varry for development, staging, production environments. eg:

   staging

   ${ENVIRONMENT}:
       host: ${REDIS_HOST}
       namespace: ${REDIS_NAMESPACE}
       port: 6379

   production:

   ${ENVIRONMENT}:
       host: ${REDIS_HOST}
       namespace: ${REDIS_NAMESPACE}
       port: 6379
       connect_timeout: ${REDIS_CONNECT_TIMEOUT}
       timeout: {REDIS_TIMEOUT}

2. Config file content:

   The content of the config files could again be populated dynamically.

we could generate config-templates or config-file-structure using helm and mount them in the required place (say config/redis.yaml ) inside the pod.Then these config-templates can be converted/rendered into proper file using utilities at container run time.

Dynamic creation of config files based on the environment/environment_variables could be achieved using following ways.

The Bash Way (using eval and cat):

1. Create a file named inator with the following content

   #!/bin/bash
   eval "cat <<EOF 
   $(<$1)
   EOF
   " | tee $1 >/dev/null

2. Make inator as executable and place it inside the docker image and execute it as a ENTRYPOINT script

3. Considering the env varibles are avilable inside the pod/container

   eg: staging

   $ env
   ENVIRONMENT=staging
   REDIS_HOST=abc.com
   REDIS_NAMESPACE=inator
   
   $ cat config/redis.yaml
   ${ENVIRONMENT}:
      host: ${REDIS_HOST}
      namespace: ${REDIS_NAMESPACE}
      port: 6379
   
   $ ./inator config/redis.yaml
   $ cat config/redis.yaml
   staging:
     host: abc.com
     namespace: inator
     port: 6379

   production

   $ env
   ENVIRONMENT=production
   REDIS_HOST=redis.prod.com
   REDIS_NAMESPACE=prod
   REDIS_CONNECT_TIMEOUT=5
   TIMEOUT=10
   
   $ cat config/redis.yaml
   ${ENVIRONMENT}:
      host: ${REDIS_HOST}
      namespace: ${REDIS_NAMESPACE}
      port: 6379
      connect_timeout: ${REDIS_CONNECT_TIMEOUT}           
      timeout: {REDIS_TIMEOUT}
   
   $ ./inator config/redis.yaml
   $ cat config/redis.yaml
   production:
     host: redis.prod.com
     namespace: prod
     port: 6379
     connect_timeout: 5
     timeout: 10

Advantages: no additional package required.

The dockerize way:

dockerize is a utility to simplify running applications in docker containers. It internally uses go templates to populate config files from environment variabels.

Have a look at this blog post Environment Variable Templates for further information.

Maybe helm can help you out here. Helm allows you to template your k8s yaml files and bundle them into one deployable k8s package. Your configuration files might be placed in ConfigMaps and can also be templated with helm. How this is done in detail can be seen in the chart template guide.

One way to fill these templates at deploy time then would be:

helm upgrade --install my-service /path/to/my/chart/package --set my.value=${MY_VALUE_ENV_VAR}