libcluster VS flyctl

Distributed Elixir can be done with Docker containers too, see https://github.com/bitwalker/libcluster which by default has some Kubernetes support but you can also have third party (or custom) clustering strategies. I've not done this myself but I've seen articles about this a lot during the past years.

Hot code updates for most applications aren't really worth it in my opinion, assuming you do something like blue/green rollover deployments. It's cool that it's possible though. But it requires appup files and afaik Distillery is one of the release tools that has support for it built-in.

On the deployment level, you can use synchronous communication using API-based solutions or asynchronous communication using external solutions (RabbitMQ, Kafka,...) or Erlang/Elixir tools (message passing, pg2, PubSub, Libcluster,...)

libcluster – Automatic cluster formation/healing for Elixir applications.

Based on the fact that your ideal is to have a similar experience to heroku than managing your own server setting up reverse proxies take a look at these options:

1) https://dokku.com - lets you turn your light sail instance basically into heroku

2) https://render.com

3) https://fly.io

4) If you have aws credits this is their heroku equivalent: https://aws.amazon.com/elasticbeanstalk

above is not what I do but would be the options I would pursue if I understand your preference and requirement correctly.

Thanks for the tips! I'll take a look - requests are typically super fast, so perhaps the database is just under high load due to launch. (I run replicas in a couple cities on https://fly.io, so general latency should be low too.)

For the filenames - I'll take a look. I'm using some popular libraries for files, which I hope address things like this.

- fly.io did a better job than Render, but I feel that it's private networking abstraction is still too complicated => it took me 3 hours to research on how to connect to Postgresql connection string, the flycast domain and sslmode=disable are really confusing => thus affect DX. Also, on Google I heard many complaints about fly.io unstable infrastructure. But on the other hand, it has the lowest API response time in my region (APAC). And the free plan is very generous for pet projects

It's running here, hosted by https://fly.io - https://www.datasette.cloud/blog/

Since we're all sharing our favorite simplest solutions, I guess I'll throw https://fly.io out there. The DX is far from perfect right now so this answer is _ever-so-slightly theoretical_ but, if you know how to use Docker, `fly launch` is extremely hard to beat.

Fly.io

Website is a bit sparse, can you explain why I'd use this over the many existing tools and services from AWS, GCP, fly.io tec?

With fly.io I can see the difference since I'm paying for compute to always be active but with their free tier its not an issue until a certain scale anyway and their paid tiers are cheap enough that it probably works out cheaper than serverless pricing

Try digital ocean app platform (not the droplets), fly.io