Ask HN: Anyone here self-hosting databases and needing advanced features?

[SirusCodes]

My friends and I are trying to understand whether a particular niche actually exists or if we’re imagining a problem that most people have already solved another way.

We’re specifically looking for people/teams who:

1. Self-host their own database (e.g., Postgres)

2. Manage it by SSH-ing into the machine and/or running automation scripts

3. Still need more advanced features: PITR, PgBouncer, observability, replication, etc.

We’re building a tool that provisions and manages database instances/clusters over SSH. Think “supercharged self-hosted DB automation,” but without requiring Kubernetes, Docker orchestration, or a full-stack PaaS. Essentially: managed Postgres-like ergonomics, but for people who want to stay on their own servers.

Right now, we’re stuck between a rock and a hard place:

1. On one side, we don’t see any product like this, except Crunchy Postgres, which is very enterprise-focused.

2. On the other, we’re not sure whether the lack of products means the market doesn’t want this, or whether engineers are just stuck doing everything manually because there isn't a lightweight option.

Current alternatives we see:

1. Kubernetes Operators – powerful, but require committing to K8s.

2. Crunchy Postgres – great, but very enterprise-oriented.

3. Full-stack deployment platforms (Coolify, Dokploy, etc.) – nice UX, but limited when you need PITR, PgBouncer, logical/physical replication, deep monitoring, etc.

4. Fully self-managed – which means a lot of SSH, scripts, cron jobs, custom backups, and playing “accidental DBA.”

We’re trying to figure out:

• Do engineers actually want a tool that handles advanced DB lifecycle management over SSH?

• Or is the expectation that if you need these advanced features, you should either use a cloud-managed DB or Kubernetes Database operators?

• If you are self-hosting databases today, how painful is the operational overhead? What does your workflow look like?

We’d love to hear from anyone who fits (or used to fit) into this category, or anyone with strong opinions on whether this is a real problem. Happy to share more details if people are interested, but mostly trying to validate whether this is a rabbit hole worth continuing down.

[kartiksoneji]

I've self-hosted postgres a few times, mostly to cut costs on dev. Two production use-cases I've come across is using custom extensions and deploying to a bare-metal compute provider.

[SirusCodes]

How did you manage it? Did you write your own script or something?

[grekowalski]

99% of small companies and startups use only simple SELECTs, INSERTs, and UPDATEs. Sometimes we also happen to do a JOIN.