Alpha and staging deployments are automatic and triggered once something is merged to master.
Production deployments should happen a few times per week, but maximum once per day. See QA for how it does affect the deployment. Specific deployment information for the applications can be found in the specific repositories either in the README or in separate documentation.
The support team of the iteration does the production deployments. Every day in the morning, the support team goes to the #deployments channel in Slack and searches for the most recent deployment proposal. If it wasn’t deployed, they click the "Deploy" button and monitor the deployment. There may be multiple deployment proposals in the #deployments channel. The most recent one should contain the highest number of QAed changes and therefore this one should be deployed, in order to avoid deploying things in two deployments unnecessarily.
If there are no deployment proposals for two consecutive days, the support team should search for the reason in coordination with people doing QA (too high QA workload, rejected tickets) and try to unblock the next deployment.
The tool that we use for orchestrating the deployments is Deployer. The README of the project contains more technical information about how the deployment workflow functions.