Open-source project
autobase-tech/autobase avatar
autobase-tech/autobase

Autobase for PostgreSQL: Ansible plus a web console, and a split between an MIT core and a proprietary console

Automated database platform for PostgreSQL® - Your own DBaaS.

4,449 stars617 forksTypeScriptMIT

At a glance

What is it?
Autobase packages Ansible playbooks and a web console into an internal DBaaS for self-hosted PostgreSQL. It is a real platform rather than a thin wrapper, but the Console is proprietary software, with even its free Starter plan governed by a separate Platform license.
Who is it for?
Adopt Autobase if your PostgreSQL estate already runs on Debian, Ubuntu, CentOS Stream, Oracle Linux, Rocky or AlmaLinux and you want one code path for deployment and failover. Do not adopt it if you need cross-distribution support, if your team wants no Ansible at all, or if you need the Console without accepting its proprietary Platform license.
Can I use it commercially?
Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
Is it still maintained?
Yes. The repository last received commits 3 days ago.
What is it written in?
Mainly TypeScript, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on November 21, 2026, and from our analysis. They are not legal advice.

From daily snapshots

Star history

  • autobase-tech/autobase4,449 ★ · +71
Star history
Repository2026-09-112026-11-21
autobase-tech/autobase4.4k4.4k

71 Stars in the last 71 days, measured from the daily GitHub snapshot. Compare with other projects

Editorial analysis

Three front doors onto the same cluster, and what each one omits

Autobase is not a database server and not a proxy. It is a control plane that stands next to PostgreSQL and drives the machines under it. The repository splits into console/ for the web UI, automation/ for the Ansible collection, and .github/ for the test workflows, and the README routes each of those to a separate entry point: the Console, the command line through Ansible, or GitOps.

The Console is where the edition split bites. Autobase Platform (Console) is proprietary software, and its use, including under the free Starter plan, is governed by a separate Platform license, while the Ansible automation core is MIT licensed.

GitOps is the third door and the one that most changes the shape of a team's workflow. Cluster configuration is kept in Git and applied through CI/CD pipelines. Nothing in the README explains merge conflict resolution, drift detection, or who wins when a human edit and a commit disagree, so treat that as an open question for your own process rather than a settled capability.

The repository's own Makefile shows the bootstrap chain

The Makefile at the repository root is the most concrete thing in the project, and it is worth reading before anything else. Its header tells you the intended order directly:

bash
make && make bootstrap

The target bootstrap chains to python-bootstrap, and the Python interpreter is not hardcoded. It is resolved by reading .config/python_version.config:

bash
python_launcher := python$(shell cat .config/python_version.config | cut -d '=' -f 2)

So the version of Python your automation uses follows a checked-in config file rather than whatever python3 happens to be on your PATH. For development there is bootstrap-dev, which calls bootstrap and then python-bootstrap-dev, and reinitialization, which runs clean and bootstrap again to return to a fresh state.

Makefile also sets .ONESHELL, .SHELLFLAGS of -eu -o pipefail -c, and MAKEFLAGS with -j and -l pinned to NUM_CORES. In practice that means one shell per recipe line group and fail-fast behaviour on the first error, which is the right default for a tool that provisions database machines.

Distribution support is a closed list, and the test table shows which entries are real

Autobase only runs on Red Hat and Debian derivatives. The README names Debian 11, 12 and 13, Ubuntu 22.04, 24.04 and 26.04, CentOS Stream 9 and 10, Oracle Linux 8, 9 and 10, Rocky Linux 8, 9 and 10, and AlmaLinux 8, 9 and 10, across x86_64 and aarch64. Anything else, including the non-stream CentOS builds that dominated enterprise fleets a decade ago, is out of scope.

The detail worth pausing on is the gap between the support list and the test table. The support list covers seventeen distribution entries, but the daily automated deployment table links workflows for twelve: Debian 12 and 13, Ubuntu 24.04 and 26.04, CentOS Stream 9 and 10, Oracle Linux 9 and 10, Rocky Linux 9 and 10, AlmaLinux 9 and 10. Debian 11, Ubuntu 22.04, and every 8-series entry appear in the compatibility list with no scheduled workflow behind them. Read that as a claim to check against your fleet rather than a guarantee.

Supported PostgreSQL major versions are listed as 10 through 18, with availability depending on the operating system and package repositories. That range is wide enough that the pin for any single cluster still has to come from you.

What the repository does not say about failure

Here is the omission that decides adoption for some teams. The README does not document rollback of a failed deployment, does not describe what happens to in-flight connections during a failover it performs, and does not state a recovery time objective for any of the automated operations it advertises. Deployment, failover, backups, restore, upgrades and scaling are all listed as automated, and the mechanics of each failure are not.

The architecture page at autobase.tech/docs/overview/architecture is pointed to for cluster components, so some of this detail lives there rather than in the repository. From the repository alone you cannot tell which of these operations run in place on a single host and which move traffic between nodes, and that difference decides whether your application can absorb a restart.

The contact address published in the README, [email protected], is offered for help implementing Autobase into an infrastructure. That is a services relationship, not documentation, and it is the practical answer to the gaps above for teams that buy support.

Ansible as the layer under the UI is what you inherit, not an optional extra

The automation/ directory is described as an Ansible collection for deploying and managing PostgreSQL clusters from the command line, and it carries its own README. The README calls this the open-source automation core, built with Ansible.

For a team already running Ansible, this is the cheapest path in. Your existing inventory files, vault handling, and CI runners apply, and the GitOps route reduces to committing cluster configuration into a repository that a pipeline already knows how to execute. For a team without Ansible experience, the README is honest that the project targets people who want the managed database experience without specialised expertise, which places the burden on the product rather than on your playbook hygiene.

Neither route removes the operational floor. Backup retention, replica placement and upgrade windows still need decisions, and the README does not enumerate defaults for any of them.

Where a plain Ansible role differs from Autobase

The obvious alternative is writing your own Ansible roles for PostgreSQL with something like the community-postgresql role and wiring HAProxy or Patroni yourself. The difference in approach is packaging and support surface. Your own roles are a single self-contained thing whose failure modes you can read end to end; Autobase is a platform that has already made choices about cluster topology, distribution pinning, and which distributions it tests daily, and it exposes those through a console and a GitOps workflow.

The cost runs the other way. Once you adopt Autobase, the topology decisions are made for you, and the Console that drives it is proprietary software under its own Platform license. A team that only needs a standby on a second host may end up installing more than it uses, while a team that needs the full set would otherwise spend months assembling it from separate roles.

Commercial hosting is the third option and it inverts the trade: no cluster management work at all, no control over the machine, and a recurring bill.

Seven years of development and two businesses in one repository

Autobase has been under development since 2019, its last push was on 2026-11-02, and the code is MIT licensed. The MIT licence covers the automation core and public deployment files; Autobase Platform (Console) is proprietary and governed by its own Platform license in console/LICENSE, including under the free Starter plan.

One contradiction is worth naming. The same README that positions Autobase as the work of a PostgreSQL expert and founder also asks readers to star the repository on GitHub to help it grow. For an infrastructure tool with a proprietary platform, the growth mechanism and the product are two different businesses, and anyone budgeting for this should know which one they are buying.

Editorial conclusion

Adopt Autobase if your PostgreSQL estate already runs on Debian, Ubuntu, CentOS Stream, Oracle Linux, Rocky or AlmaLinux and you want one code path for deployment and failover. Do not adopt it if you need cross-distribution support, if your team wants no Ansible at all, or if you need the Console without accepting its proprietary Platform license. Verify first that your target distribution appears in the daily deployment test table, which lists twelve distributions, and that your PostgreSQL major is among the supported 10 through 18.

Frequently asked questions

What is Autobase?

Autobase is an internal database platform for PostgreSQL that brings a managed DBaaS experience into your own infrastructure. You create and manage production ready, highly available PostgreSQL clusters, automating deployment, failover, backups, restore, upgrades and scaling.

What is an alternative to Autobase?

The practical alternative is assembling your own Ansible roles for PostgreSQL together with a separate HAProxy or Patroni setup. That gives you one self-contained codebase whose failure modes you can read end to end, while Autobase packages topology and distribution choices into a platform with a console and a GitOps workflow.

Autobase versus what?

Autobase is usually compared against hand-written Ansible automation or a commercial managed PostgreSQL service. Autobase supplies cluster topology, distribution pinning and daily deployment tests that you would otherwise assemble yourself; commercial hosting removes the management work but takes away control of the machine.

Official sources

  1. autobase-tech/autobase on GitHub
  2. License: MIT
  3. Project website
  4. README
  5. Releases
Add this badge to your README

If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.

Add this badge to your README

markdown
[![Hysen Labs](https://hysenlabs.com/badge/autobase-tech-autobase.svg)](https://hysenlabs.com/projects/autobase-tech-autobase)

Star history chart for your README

markdown
[![Star history](https://hysenlabs.com/chart/autobase-tech-autobase.svg)](https://hysenlabs.com/projects/autobase-tech-autobase)

Was this useful?