Any build starts with a change pushed to GitHub. Since Cirrus CI is a GitHub Application a webhook event will be triggered by GitHub. From the webhook event Cirrus CI will parse a Git branch and a particular SHA for the change. Based on parsed information a new build will be created.
After build creation Cirrus CI will use GitHub APIs to download a content of
.cirrus.yml file for the SHA. Cirrus CI
will parse it and create corresponding tasks defined in the configuration file.
These tasks defined in
.cirrus.yml file will be dispatched within Cirrus CI to different services responsible for scheduling on
supported computing service. A scheduling service will use appropriate APIs to
schedule a VM instance or a Docker container on the particular computing service. The scheduling service will also
configure start-up script that downloads Cirrus CI agent, configures and starts it. Cirrus CI agent is a self-contained
executable written in Go which means it can be executed anywhere.
Cirrus CI agent will request commands to execute for a particular task and will stream back logs, caches and exit codes of the commands upon execution.
Once task finishes the same scheduling service will clean up a VM or a container.
Image above is a diagram of how Cirrus CI schedules a task on Google Cloud Platform. Blue arrows represent API calls and green arrow represents unidirectional communication between an agent inside a VM or a container and Cirrus CI.
Other things like health checking of the agent and GitHub status reporting are happening at same time as a task is running
but the main flow was described above. Straight forward and nothing magical.
For any question please use official support channels.