Why are we doing this?
While we have a reasonable setup for local development and running the OpenCodelists server, this does not entirely replicate how it runs in the production.
We don't have a good way of testing out changes that affect the Dokku deployment.
In certain cases, there may be changes that:
- either behave differently in the Dokku deployment compared with a setup without Dokku
- are not testable without having Dokku running (for example, changes to health checks)
How will we know when it's done?
When we have a working easy-to-reproduce setup that incorporates the OpenCodelists server running within Dokku, where we can rebuild the production Docker image, and see the deployment update.
What are we doing?
Possibly adding configuration that allows for local testing with Dokku?
Defining delivery tasks guidance
Why are we doing this?
While we have a reasonable setup for local development and running the OpenCodelists server, this does not entirely replicate how it runs in the production.
We don't have a good way of testing out changes that affect the Dokku deployment.
In certain cases, there may be changes that:
How will we know when it's done?
When we have a working easy-to-reproduce setup that incorporates the OpenCodelists server running within Dokku, where we can rebuild the production Docker image, and see the deployment update.
What are we doing?
Possibly adding configuration that allows for local testing with Dokku?
Defining delivery tasks guidance