I'm working on a project that's using Heroku's preboot for zero downtime. As far as I can tell they don't have an easy way to indicate when it's done with prebooting, so I've decided to implement my own custom healthcheck that will report the commit hash that it's currently running on.

Unfortunately unless I set an incredibly long delay, I won't be able to use the healthcheck option provided here because I have no guarantee of when the preboot has switched over. This would be a lot easier if I had an option to retry the health check some number of times, with some delay between them.

Alternatively this would be a ton easier if there was some way to have the step only advance when the preboot has flipped over!