Add a retry mechanism for communicating with the Foojay back-end
See this issue and related PR.
It would be good if communication failures with the Foojay back-end wouldn't immediately make the plugin's resolving fail. It would be good if the plugin would retry a couple of times before giving up. I'm thinking 2-3 retries with increasing delays, not too long though, the Gradle build might have other toolchain resolvers configured, at some point we need to hand things over to those.
Yeah, it’s frustrating that the metadata queries break the build when the APIs is down and auto-download is disabled. The resolver shouldn’t assume uptime guarantees that isn’t being provided.
The issue is in the backlog of the relevant team and is prioritized by them.