type of deployment environemt

  • Production

On what3words.comarrow-up-right or w3w.coarrow-up-right domains. Using our "legacy cluster" setup, which means the deployment process is more manual. To deploy, update the relevant Version entry for your application in the relevant file (usually this onearrow-up-right), create a PR with changes and assign to the usual devs and merge to master when approved, then run the deployment commands:

> source metabuild/venv/bin/activate # to enter the virtual environment

> ./kubernetes/helm_wrapper prod london APPLICATION_NAME # the actual deployment

2. Preprod

On preprod.w3w.ioarrow-up-right or london.preprod.w3w.ioarrow-up-right domains. This is where QA do a final regression test before releasing to production. Also a "legacy cluster" environment, so the steps to deploy are similar to preprod. In this scenario we update a different filearrow-up-right and run similar commands:

> source metabuild/venv/bin/activate # to enter the virtual environment

> ./kubernetes/helm_wrapper preprod london APPLICATION_NAME # the actual deployment

3. Non-production/dev (also called NPD)

Available on dev.non-production.w3w.ioarrow-up-right domain. This is where developers (and sometimes QA) check new changes. This is the first of our new clusters, and the setup will eventually be replicated to create a dedicated QA environment and also replace production. This deployment process is more automated, where a successful build in circleci will eventually trigger a job like in the screenshot, where all a user has to do is click the approve button. This triggers a change to the infrastructure-manifestsarrow-up-right repo which will automatically apply the changes if there are no errors, and the new version should be running in about 5 minutes.

Last updated