AWS::SSM::MaintenanceWindowTask-TaskInvocationParameters-MaintenanceWindowRunCommandParameters-CloudWatchOutputConfig
AWS::SSM::MaintenanceWindowTask-TaskInvocationParameters-MaintenanceWindowRunCommandParameters does not support the CloudWatchOutputConfig attribute documented in the API and CLI documentation and appears to only support logging to S3, not CloudWatch.
It should be possible to create a Task that logs to CloudWatch
This was a fairly surprising omission for me that cost me an afternoon and a lot of hair-pulling.
This probably seems like a corner case, until you have a number of unreliable maintenance window jobs where S3 logging won't suffice since it doesn't log things that time out. And then it's a major shortcoming. :(
And given this is available it seems from the CLI, very frustrating!
https://awscli.amazonaws.com/v2/documentation/api/latest/reference/ssm/register-task-with-maintenance-window.html
https://docs.aws.amazon.com/systems-manager/latest/userguide/sysman-rc-setting-up-cwlogs.html#sysman-rc-setting-up-cwlogs-send
+1 keen on this.
I need cloudwatch logging for my tasks and hate having to turn stuff on using CLI or through my browser as I'll forget about it when I deploy the stack to a new region
+1 on this one too
+1 on this
+1 on this
+1 on this
@cleorun @usbulat @progerjkd @1robroos
If you react with the 馃憤 button to the original issue, (the first comment, click on the smiley face if you're the first reacting), your votes can be used to sort issues and determine priorities.
A comment will send a notification to everyone (participants and watchers), but cannot be easily counted as a vote for an issue. Thus It's generally better to vote than to comment with "+1". To keep up to date, you can also add yourself as a watcher.
+1
@1robroos Can you please re-read the comment right between your two comments?
Every comment on an issue sends a notification to somewhere between 9 and 160 people (and now they got one from me too). I'd appreciate seeing fewer "+1"s in my inbox.
Most helpful comment
This was a fairly surprising omission for me that cost me an afternoon and a lot of hair-pulling.
This probably seems like a corner case, until you have a number of unreliable maintenance window jobs where S3 logging won't suffice since it doesn't log things that time out. And then it's a major shortcoming. :(