I’ve been exploring a variety of Postgres managed solutions lately and have wondered what are the considerations fly.io is making for onboarding one(or many?) managed Postgres providers or if there’s anything being done so that third party providers can integrate better. I see from this post Future of Postgresql on Fly? especially this quoted part
which mentions that the long term plan is to provide managed Postgres solutions to users similar to how Fly.io integrated Upstash for Managed Redis (great move btw). Personally I can say I have been very happy with NeonDB considering its architectural decisions such as separating compute and storage which allows for cost efficient branching/zero config rollbacks to any point in time. Additionally the Neon setup is serverless much like Upstash is for Redis, while also being Open Source and having a phenomenal support system (even as currently “small” users the email support has been truly amazing, within minutes - I promise no paid shilling here, just speak from experience).
Neon’s ability to scale to 0 but also horizontally scale to any need makes it a perfect fit for the needs of a growing startup (cost/resource conscious) or one that is well established. Additionally from what I can see on their documentation/newsletters they are working on adding cross regional read replicas which slots in nicely with Fly’s multiregional focus. Right now, as mentioned in other posts, there are some drawbacks with using any other managed service outside of Fly.io (Neon or others) due to issues with egress/latency and security so having Neon as an “internal” managed option would be really amazing.
If it helps to build the case at all, Neon has already integrated with Vercel Neon Postgres – Vercel & Neon and Vercel overview - Neon Docs so I don’t think they would be any strangers to integration.
Would be curious to know if there has been any contact or if there’s any progress on this side of things? Really appreciate it!
Other posts on this topic: Plans for fully-managed postgres?