Introduction
Sometimes, when you deploy an AWS ECS service using CDK or a CloudFormation stack, you end up in a strange situation where the deployment seems to go on forever. The tasks aren’t deployed, the tasks are failing health checks, and ECS just keeps starting new tasks.
Sometimes, you can see even the old tasks aren’t passing health checks. That’s a guarantee that your poor service is going to take forever to fail, according to CloudFormation.
You’re waiting and waiting and waiting. You already know how to fix it and are ready to move on; however, CloudFormation just keeps trying in vain. This can be a really frustrating experience.
How to Trick CloudFormation Into ‘Thinking’ It Succeeded
So, a basic way to trick CloudFormation into thinking everything has been deployed successfully is to redeploy the service in the AWS Console.
Yes, that’s right. You can redeploy the service in the AWS Console, but this time, you are going to set the following parameters:
- DesiredCount: 0
- MinCount: 0
- MaxCount: 0
REMEMBER: You are not changing anything else in the service; you are just changing the desired, min, and max count to 0. Also, you don’t need to check the box to force a redeployment.
After all the tasks have been stopped, CloudFormation will think that everything has been deployed successfully.
This will allow you to continue debugging your service without waiting hours for CloudFormation to time out or for the number of retries to finally be exhausted.
This trick has saved me much time and frustration over the years.
Comments are closed