You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The CI utilizes the official test runner images published by GitHub. The problem here is that GitHub runners get updated frequently with dependencies that prevent our application from starting without us doing extra steps, due to the nature of the project being able to support multiple versions and avoid backwards incompatibility.
We also do not actively track releases and sometimes the release notes for the github runner images are not updated to accurately reflect the actual versions in the runners. This causes errors in our CI that we were not expecting and causes blockers on PRs until we allocate resources to fix the CI problems.
Describe the solution you'd like
Utilizing our Dockerfile we create a GitHub specific Dockerfile and publish an image potentially with a generic entrypoint to install dependencies. But ideally a static Dockerfile that limits the amount of random issues to address for example, we do not have to worry about the chrome version changing.
We can also reduce the runtime of our CI by reducing the setup steps. The CI runs through setup steps like setting up yarn and NVM. This could be just be built into the Docker image and then ensure the GitHub user code 1001 has access to these folders so that these setup steps can just be skipped.
However, we will have to maintain this Docker image which is another responsibility but since we maintain it then we do not have to worry about an upstream resource breaking our CI. We should also consider having an un-official docker hub to have full control on this image.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
Origin: #3997
The CI utilizes the official test runner images published by GitHub. The problem here is that GitHub runners get updated frequently with dependencies that prevent our application from starting without us doing extra steps, due to the nature of the project being able to support multiple versions and avoid backwards incompatibility.
We also do not actively track releases and sometimes the release notes for the github runner images are not updated to accurately reflect the actual versions in the runners. This causes errors in our CI that we were not expecting and causes blockers on PRs until we allocate resources to fix the CI problems.
Describe the solution you'd like
Utilizing our Dockerfile we create a GitHub specific Dockerfile and publish an image potentially with a generic entrypoint to install dependencies. But ideally a static Dockerfile that limits the amount of random issues to address for example, we do not have to worry about the chrome version changing.
We can also reduce the runtime of our CI by reducing the setup steps. The CI runs through setup steps like setting up yarn and NVM. This could be just be built into the Docker image and then ensure the GitHub user code
1001
has access to these folders so that these setup steps can just be skipped.However, we will have to maintain this Docker image which is another responsibility but since we maintain it then we do not have to worry about an upstream resource breaking our CI. We should also consider having an un-official docker hub to have full control on this image.
The text was updated successfully, but these errors were encountered: