I have a scaling group (ASG) which I use to keep instances ready for a task. Instance is detached from the ASG when needed and ASG starts another instance when needed.
The instance type is c6g.4xlarge. OS is ubuntu 22.04
Below is the log showing the startup time.
systemd[1]: Startup finished in 4.850s (kernel) + 42min 16.004s (userspace) = 42min 20.854s.
I can see multiple of these failing SSM logs
ERROR [Registrar] failed to register identity: error calling RegisterManagedInstance API: RequestError: send request failed
caused by: Post "https://ssm.us-xxxx-x.amazonaws.com/": dial tcp xx.xx.xx.xx:xx: i/o timeout
INFO [Registrar] sleeping for 18.7 minutes before retrying registration
This is happening for only a few percent of instances and I am not able to figure out a pattern.