autoscaling waiters

[aleksandersumowski]

I'd like to script a blue-green deployment between two auto scaling groups.

Having a waiter that would tell me when instances in the new group have finished launching would help a lot.

[mtdowling]

Thanks, I'll mark this as a feature request. In the meantime, I think you'd need to describe the autoscaling group and poll the instances in the group until they are marked as InService.

[kyleknap]

+1 on feature request from user who contacted me

[quiver]

+1 (I need this feature in my project)

aws/aws-sdk-ruby has already implemented this but has some bugs aws/aws-sdk-ruby#1109 .
Once it's fixed, just copy the waiter JSON to botocore and it should work fine 😃

[kyleknap]

@quiver I will pull the waiters from the ruby sdk once they get fixed.

[quiver]

@kyleknap thx.

[revmischa]

+1 definitely need this for https://github.com/revmischa/udo

[fzk-rec]

Any update on this?

[gyoza]

This would be great for a project we're doing at the company I work at also. Right now I just have a 30 minute celery job that checks to see if anything has been done ...

[fzk-rec]

My workaround is to check every 30 seconds if the ASG still exists:
response = client.describe_auto_scaling_groups(...)
while response['AutoScalingGroups']:
time.sleep(30)

[gyoza]

That type of a solution wouldn't be ideal when your AWS account is getting close to its API limits. Loops and whiles on response codes is a solution but its a bit dirty and could be nicer.

[dmulter]

Today I'm polling, but would much prefer waiters for the following:

• ASG deleted
• At least one EC2 instance has started in a new ASG

[ismaelfernandezscmspain]

+1 (I need this feature )

[gyoza]

Just a thought for everybody 'polling' -- could always just monitor SQS for messages from the ASGs in the meanwhile.

[yks0000]

I was looking for waiters for Attach and Detach ELB in ASG. Searched the boto3 docs but did not see any waiter name there.

Does any one implemented it or any suggestion in this regards.

http://boto3.readthedocs.io/en/latest/reference/services/autoscaling.html#AutoScaling.Client.get_waiter

[Gaafar1]

is there any waiters for reinstatement for New instances in ASG to come on air if old instance is unhealthy or terminated ?

[ismferd]

Have you tried with cfn-helpers? cfn-signal could help you.
https://docs.aws.amazon.com/es_es/AWSCloudFormation/latest/UserGuide/cfn-helper-scripts-reference.html

[cognitiaclaeves]

I wandered into this forum while looking for a waiter for ASG. As of this date, there still isn't one, but I noticed something in this thread that I found misleading. There is an implication that waiters are solutions that do not poll, and therefore are preferable to writing your own poller. I don't believe this is the case. Here is what the documentation says about a waiter (for ec2):

wait(**kwargs)

Polls EC2.Client.describe_instances() every 15 seconds until a successful state is reached. An error is returned after 40 failed checks.

( source: https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/ec2.html#EC2.Waiter.InstanceStopped )

So I think that rolling your own poller should be fine. ... For anyone that is waiting on a waiter for a particular service, I also found out that you could implement your own:

https://www.2ndwatch.com/blog/use-waiters-boto3-write/

Also, the way the existing waiters are implemented can be found in botocore and used for examples of how to write your own:

https://github.com/boto/botocore/blob/master/botocore/data/ec2/2016-11-15/waiters-2.json

[tim-finnigan]

Thanks all for your patience. Because service teams now own the creation of their waiter models for use across SDKs, I created a new tracking issue for this in our cross-SDK repository: aws/aws-sdk#419. And I reached out to the AutoScaling team to see if they would consider adding the waiter models requested in this issue.

If any one has any additional feedback on this request please let us know and you can refer to that new issue for updates going forward.