Remote Deployments Instable


I’ve noticed that remote builder deployments have always been stable. Almost every other deployment will error with:

Error error connecting to docker: error establishing wireguard connection for xxx organization: error fetching dialer: establish failed: err err handling establish: can't get wireguard state for xxx: An unknown error occured.

Sometimes it will be another error bbout context or something like that

“An unknown error occurred” is an api error. We’ll look into it.

The other timeouts are a huge problem for remote builders and ssh in general… Here’s an overview of the issue: WireGuard and Remote Builder Fixes in flyctl - #2 by thomas. The fix is moving remote builders off nomad/consul and onto our upcoming microvm api. You can try it out now by running flyctl deploy --remote-machine-only in the latest flyctl prerelease.

curl -L | sh -s pre

We’re hoping to share more on this soon.

I found your error in Sentry. I’ll get a fix out tomorrow morning. Sorry about this!