Hi folks. There are a collection of issues in this repo describing an increasingly escalating situation where this GitHub Action is nearly unusable due to high latency on new tailnet nodes. I've tried and failed to deploy a package update to my homelab twice in the last day, so I thought I'd open this issue ahead of exploring a new deployment approach this weekend.

Setup takes ~roughly the duration of the timeout and sometimes fails

• Setup Tailscale step taking two minutes #159
• Setup step is failing in first run across multiple repos #177

First connection after successful setup sometimes fails

• Intermittent connection issue after connecting to the remote host using tailscale #145
• Urgent: intermittent connection to remote host through Tailscale action #176

#177 (comment) contains an acknowledgement of the platform issues at play here. Does the Tailscale team have any updates to share?

In the presence of the latency described in the above issue, there's a second-order at play here: sometimes Tailscale setup reports that setup completes successfully when the latency on the connection is so high that subsequent connections are likely to fail. In the presence of endemic latency during this phase of connectivity, is there a way that this action can detect this latency condition and wait until it settles before returning success or is that an exercise left up to the reader? 😅