S3 returns an xml document when the requested file does not exist, meaning curl will still exit zero and testing for the curl exit code is worthless.

One option, file detects the file as xml, when it should be gzip compressed data or zip archive data.

Would be nice to have the aws utility on all RDSH instances, and it does requires admin privs to install so it is not something users can install to their own profile.

• https://s3.amazonaws.com/aws-cli/AWSCLI64.msi

Some experimentation may be needed to figure out the quiet install flags, but it probably looks much like the chrome install...

• https://github.com/plus3it/cfn/blob/master/scripts/configure-rdsh.ps1#L92-L97

• Set/increase the idle timeout to 900 seconds for the RDGW and RDSH ELBs
• Enable connection draining with an 1800 second timeout for the RDGW and RDSH ELBs
• Use scheduled actions control how many Guac, RDGW, and RDSH instances are running
  • Think about how to parameterize the scheduled action