We have a heroku review app pipeline configured with github integration.
Currently, every time after a new app is created, we have to go in and manually disable automatic deploys in the UI here:
I was looking around for one of the following options to remove this manual step:
call an api to disable auto-deploy on a review app
change a pipeline level setting that determines the default value of auto-deploy for new review apps that get created
but so far I haven't been able to find anything.
Any suggestions for how this might be achieved would be much appreciated.
I've installed the Google Cloud Build app from Github Marketplace, and enabled it on several repositories, only to discover that enabling access for those repositories automatically starts running the Google Cloud Build "check" after every PR or push to a PR. Since some of those repositories are not actually ready for Google Cloud build, I needed to turn off the checks... I used the URL mentioned in the docs to get to the page where you can select which repositories should be connected to your GCP project, and unchecked a bunch of them... this disconnected them from the project, but, surprisingly, did not disable the Github checks! So now, there are Github checks on several repositories that just result in an error because they're not connected to any projects in Cloud Build.
Surely, it is not an uncommon use case to need a way to "disable Checks" on a per-repository basis - is there any way to do that?
(If not, consider it a requested feature if there are any Google Engineers reading!) ;)
I did talk to chat support, and the only way we could figure out to remove the checks would be to uninstall Google Cloud Build completely from the entire organization, which I don't want to do because we do have some setup that I want to keep, and I don't know how much of that configuration would have to be redone if I uninstalled it completely.
Here is a screenshot of the broken check that I can't disable:
I think there are 2 steps to disable CloudBuild,
1. Disable on CloudBuild itself. For some time I kept seeing 2 builds per push made, but disabling a trigger on CloudBuild helped fix that.
2. On Github, you can configure the Cloudbuild app to monitor a specific repo on the organization level setting.
3. If you feel a lot adventurous though and it's not solved, on the project level setting, you could delete the Webhooks integration for cloudbuild, as that's what Github notifies of each build.
It can be found within GH App installations:
https://console.cloud.google.com/gcb-github-registration
Select the GH account used for your link between Cloud Build and GH, then the GCP project you want to edit.
Here connected repositories can be edited, and in the second step (Trigger settings), the "Default GitHub Checks" can be disabled.
You can also goto settings of your repository and click installations inside settings.
And then remove Google cloud build from Installed apps and authorized github apps.
Cloud Build Triggers are still in beta, and have gone through several updates since this question was posted, but it now appears that the "Github Default Checks" are no longer hidden from management. They are now explicitly offered (but not automatically created) when connecting a repo, and you can finally view and disable (or delete) individual "Github Default Checks" on the trigger management page: https://console.cloud.google.com/cloud-build/triggers
I'll leave this as unanswered, because I'm uncertain this is how it is working for everyone, or that this answer will remain true, but if it works for you, feel free to vote up the answer!
To link GCP Project to the Github repository, create a trigger by clicking on resolve link, besides Action Required and then disable the same. For me this procedure worked out.
We are using Github + CircleCI + Heroku with automatic deploys setup.
From time to time there are deploy errors that are not captured by CI (deployment fails event though CI run is fine). We have review apps set up for all our PRs.
It would be great if the PR would indicate this and stop us from merging if the deployment of the review app fails (that usually means the deployment of the staging/production app would also fail).
I could not find any documentation on this. The only possible way I see is to use the GitHub API to add a custom check for this.
Anyone solved this issue?
We have made a GitHub Action to test the deployment status of a Heroku Review App: https://github.com/marketplace/actions/heroku-review-app-deployment-status.
This can be used in the GitHub workflow to test the deployment status.
UPDATE: We wrote a nice blog post about different approaches to verify Review App deployment status.
Link: https://blog.niteo.co/staging-like-its-2020/
You can use Github's protected branches for this https://help.github.com/en/articles/enabling-required-status-checks
Navigate to your repository's settings -> Branches -> Add Protection Rule -> Select "Require Status Checks Before Merging" and select "CircleCI".
In order for this to work you need to cause CI to fail if your deployment fails. If you get a failure from Heroku, you should run any command that will return exit 1 as a status code, which will fail CI for you.
Why am I getting this error message for my apps deployed with Github at Heroku?
There is an issue with the GitHub token for this app. Disconnect and reconnect to restore functionality
We had the same issue before, it happened with private repositories and disconnecting/reconnecting from time to time seems to do the trick but after a while we started to grow and we needed more automation.
We looked into all kinds of CI/CD tools like Codeship, CircleCi, etc. Ended up choosing DeployBot and stuck with it because it works really well for us and fitted our needs best.
Either way these tools can be lifesavers for the team no matter which one you end up using
For whatever range of poor reasons...a user schedule workflow has been uploaded to our YouTrack server instance that now automatically generates seemingly hundreds of new issues a minute.
As you might imagine, the web server is totally unresponsive, meaning that we cannot disable or detach the workflow from the project. The workflow editor's attempts to remove the workflow fail with connection timeouts.
How can the workflow be disabled without accessing the usual admin pages? Is there a workflow file somewhere in the server's directory that can be deleted (I can't find one) while the service is stopped? Is there a "run in safe-mode" option for YouTrack?
We're running YouTrack 5 fwiw.
Thanks in advance.
In YouTrack 6.0 we've implemented special Java start parameter (-Djetbrains.youtrack.workflow.detachModified=true), it will work starting from next bug fix release.
This option will detach all the workflows from all the projects on start, except the unmodified supplied ones.
Please, consider https://youtrack.jetbrains.com/issue/JT-27785, the Fix version of this issue will be changes as soon as we release next bug fix.
Workflows are stored inside YT db, so the best variant for you is to contact YT support. It'll be able to modify the db.