My script: cml runner launch \ --repo=... \ --labels=cml-cloud --single --idle-timeout=120s \ --cloud=aws --cloud-region=us-east-1 --cloud-type=c6a.8xlarge
It seems to ignore --idle-timeout on AWS, while it worked well on GCP. The result is that my instances keep running forever and eating money.
My script:
cml runner launch \ --repo=... \ --labels=cml-cloud --single --idle-timeout=120s \ --cloud=aws --cloud-region=us-east-1 --cloud-type=c6a.8xlargeIt seems to ignore
--idle-timeouton AWS, while it worked well on GCP. The result is that my instances keep running forever and eating money.