v20240724-192912

Fix bug with emitting logs from runners (#5503)

v20240724-185114

On ScaleDown remove old/stale SSM parameters (#5479)

v20240724-160740

Small improvements designed to increase the resiliency of ephemeral i…

v20240719-185908

Emit Runner info at the start of every job on self hosted linux runne…

v20240718-180515

Remove workaround to install kernel-devel (#5480)

This workaround might not be necessary any longer as kernel-devel
appears to be available again in package managers. This also fixes the
LF ALI runners as they are failing to fetch from the s3://ossci-linux
bucket due to no permissions.

I verified by doing a `yum install kernel-devel` on one of the EC2
instances that was stuck and saw that it returned a newer kernel-devel
package.

Relates to pytorch/pytorch#129880 and pytorch/ci-infra#244.

Signed-off-by: Thanh Ha <[email protected]>

v20240715-170934

Adds metada options on launch instance templates to set idsv1 optiona…

v20240712-180343

Move user-data and daemon monitoring to a separate CW namespace (#5424)

v20240712-140548

Improves visibility for AMI experiments (#5421)

v20240710-170951

Adds support to experiment with AMI runner-by-runner (#5406)

v20240704-191103

Userdata send cloudwatch metrics to setup alerts on failures (#5399)