Runner lambdas v20230131-205135

limit messages delay to  900s on retry (#1966)

Runner lambdas v20230131-135103

creates a sqs queue for retry failed scale-up requests that are retry…

Runner lambdas v20230202-222408

infra: Remove hardcoded ephemeral flag, use config value

Windows were hardcoded to be ephemeral but we need the ability to
actually set these to be non-ephemeral

Signed-off-by: Eli Uriegas <eliuriegasfb.com>

ghstack-source-id: 4dfacd11f858b53edb9ed38cfdf9c80c43369e45
Pull Request resolved: https://github.com/pytorch/test-infra/pull/1420

Runner lambdas v20230111-150241

fail scale-up batch when a single instance of the batch fails (#1384)

Runner lambdas v20230111-135937

improving retry dynamic for scale-up sqs because of the ephemeral run…

Runner lambdas v20230106-151153

Fix edge cases on #938 'Fix runaway runner deletion on scale-down whe…

Runner lambdas v20221220-141244

manage mac pet instances with terraform (#1268)

Runner lambdas v20221125-134255

Remove unused VPC and SecGroup after layout migration (#1185)

Runner lambdas v20221124-080939

Fixes to correctly assign the security group to the ec2 runner (#1144)

AWS security groups need to be assigned to the correct VPC, with the new
design of multiple VPCs there is a need to manage multiple SG and the
hierarchical relationship for Region -> VPCs -> (subnet, sg);

Runner lambdas v20221118-104528

Jeanschmidt - Add partially multi-region logic && separate VPCs for c…