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
My development team at work uses flux/flagger for all of our deployments. The biggest problem we have is the fact that our application is a monolithic monorepo. We deploy our backend and frontend (sometimes multiple front ends to one backend) at the same time, with multiple canary's and AC tests and whatnot.
However, we want to the frontends to continue in canary analysis if the backend has successfully deployed. If the backend fails, it should fail the dependent canary's.
Proposed solution
webhooks that only execute open canary success or canary failure
some sort of field within canary object that allows for them to communicate with each other?
Any alternatives you've considered?
manual gating. We do not want manual processes
separate charts into multiple charts, that then use api requests to GitHub to bump the frontend chart versions in order to trigger the deployment process (current implementation).
Are there any other solutions? I spent a considerable amount of time looking at the gates, and they are close to being a fit but not quite due to lack of support for webhooks that only execute on success or only on failure. Also they can be convoluted. We don't want to have our actests sending requests to the flagger deployment if we can avoid it.
Looking for other suggestions, thanks!
The text was updated successfully, but these errors were encountered:
Describe the feature
My development team at work uses flux/flagger for all of our deployments. The biggest problem we have is the fact that our application is a monolithic monorepo. We deploy our backend and frontend (sometimes multiple front ends to one backend) at the same time, with multiple canary's and AC tests and whatnot.
However, we want to the frontends to continue in canary analysis if the backend has successfully deployed. If the backend fails, it should fail the dependent canary's.
Proposed solution
Any alternatives you've considered?
Are there any other solutions? I spent a considerable amount of time looking at the gates, and they are close to being a fit but not quite due to lack of support for webhooks that only execute on success or only on failure. Also they can be convoluted. We don't want to have our actests sending requests to the flagger deployment if we can avoid it.
Looking for other suggestions, thanks!
The text was updated successfully, but these errors were encountered: