Glad you're finding it useful! Have you set 'TLS > Origin Server Name' correctly when you configured the tunnel routing?

I'm not familiar with JAMstack, but you might be able to run separate instances in all your different locations, each with its own Cloudflare tunnel. CF allows multiple tunnels running at once, so perhaps that handles load balancing for you?

No problem! Yes, my Swarm nodes are all at the same physical location. I think you're right that dispersing nodes could lead to issues, but depending on what services you're running, you might be able to run separate Swarms with some kind of fail over?

No problem! Yes, the `cf_token` secret is read directly, so is required - github.com/tiredofit/do...

You're right, it doesn't look like `/api` is needed on the end of TRAEFIK_POLL_URL. I've got it included in my config though, and it runs fine regardless!

Unfortunately there's no way of knowing what your Post ID is before you make the Bluesky Post, so you're right - you need to go back and update the Hugo post front matter afterwards.

Glad the writeup was useful!

Thanks very much!

The PDS is so far only using a few MB of storage, I'm assuming it will only grow significantly with large blob files. Traffic is also negligible, presumably because of my low follower count!

Thank you so much for writing this up!

I've made some tweaks for my specific use case and added some additional functionality, with the resulting code available at github.com/mattdy/blues... - I've linked back to yourself as the original inspiration, hope that's okay!

No problem, glad you found it useful! I'm working on adding Bluesky as a source for my blog comments - you've got the dubious honour of being the test subject!