GSoC 2025

This page is for collecting ideas for Google Summer of Code 2025 projects. This page is for IDEAS ONLY.

IF YOU ARE A CONTRIBUTOR: there is a top-level GSoC page for PostgreSQL here: PostgreSQL General GSoC Page - please read this first before proceeding to contact mentors! Contribution guidelines, channels and communication methods are in this page too. PLEASE make sure you have read everything thoroughly.

If you are a mentor and would like to participate, feel free to edit this page and put your name below.

Mentors mailing list for proposals: gsoc2025-mentors@lists.postgresql.org

Proposals

Proposals are still work in progress. If you are a mentor and would like to participate, feel free to edit this page and put your name below.

Admins

Jesper Pedersen <jesper (dot) pedersen (at) comcast (dot) net>
Pavlo Golub <pavlo (dot) golub (at) gmail (dot) com>

Mentors

The following individuals have been listed as possible mentors on the below projects, and/or offered to be mentors for student-proposed projects:

Pavlo Golub
Jesper Pedersen
David Wheeler
Mark Wong
Akshat Jaimini
Saurav Pal
Shahryar Soltanpour
Haoran Zhang
Luca Ferrari
Rajiv Harlalka

<PROJECT NAME GOES HERE>

Project Description

Skills needed

...
...
PostgreSQL

Difficulty level

Easy, Moderate, Hard

Project size

Short (~90 hours), Medium (~175 hours), Long (~350 hours)

Consider what the length would be for the student - not the time that you can do the project in

Mentors

Person 1 -- Must be unique
Person 2 -- Can co-mentor multiple projects

Add email addresses in case pre-projects are required.

Expected outcomes

Working implementation of ...

References

link 1
link 2
PostgreSQL

Enhancements to PgWatch v3 RPC integration

Project Description

In GSoC 2024, we integrated Remote Procedure Calls within pgwatch v3 to provide a functionality called Remote Sinks. In 2025 we would like to enhance this implementation by adding more features and utilizing the full capabilities of Remote Procedure Calls.

Skills needed

Go
PostgreSQL

Difficulty level

Easy

Project size

Medium (175 hours)

Can be modified according to proposed solution

Mentors

Akshat Jaimini
Pavlo Golub

Expected outcomes

Provides an authentication protocol for the current RPC implementation
Optimized architecture

This is not a strict & exhaustive list and we encourage contributors to provide their own creative input as well

References

pgwatch v3: [1]
Remote Sinks implementation for pgwatch: [2]

ABI Compliance Checker

Project Description

Develop and deploy an "ABI Compliance Checker", to be integrated into the PostgreSQL development process similar to coverage.postgresql.org. The checker should run on every commit to the project and produce an ABI compliance report, and trigger an alert when an ABI has changed in an incompatible way. See this thread for an example of such a report, and some details on how it was implemented. Use it as starting point for building the tool.

Once the tool reliably produces reports, work with the infrastructure team to get it added to the development pipeline and to publicly publish its reports.

Background

PostgreSQL recently added Server API and ABI Stability Guidance to help extension authors to understand how and when the server API and ABI are and are not likely to change. In practice, the guidance is that the API and ABI should maintain compatibility between minor releases (e.g., 17.0 to 17.1), but not major releases. This has long been the implicit guidance, but now it is explicit, and due to be included in the PostgreSQL 18 docs.

Today the committers adhere to this policy purely through the review process, which means once in a while an incompatible change will be included in a minor release. Such changes have been extremely rare, but this past fall an ABI change was unexpectedly shipped in PostgreSQL 17.1. It was quickly reversed in PostgreSQL 17.2, but highlights the need for some process to catch such changes before they're released.

This project will help reduce the change of such an incident again by automatically checking for ABI changes. This will improve the adherence to the guidance, perhaps allow it to be upgraded to a reliability _policy_, and give extension developers and package maintainers assurance about the reliability of extension builds on PostgreSQL minor releases.

Skills needed

C Programming
Command-line tooling
HTML & CSS
Service deployment
Automation

Difficulty level

Medium

Project size

Medium (175 hours)

Mentors

David Wheeler

Expected outcomes

Working implementation of a subdomain of postgresql.org or as part of the build farm featuring ABI compatibility reports for every commit pushed to a back branch of PostgreSQL
Notifications of failures sent via email to the committers
Git repository for the project for ongoing maintenance
Integration into the PostgreSQL infrastructure build and deploy processes

References

abi-dumper, a potential tool
PostgreSQL Build Farm
Build Farm Server Code
pgsql-hackers discussion of an example of a report and the need for a project like this
Server API and ABI Stability Guidance

pg_top data internal data structure refactor

Project Description

pg_top is a terminal tool to monitor the top running PostgreSQL processes. This project is to port and use libbsd's red-black tree data structure.

pg_top doesn't need all of what libbsd has to offer, so it makes more sense to just port tree.h.

If time permits, additional tasks can be added to improve pg_top.

Background

pg_top currently uses libbsd's red-black tree data structure to organize process information but this library is not easily used on some platforms, like MacOS. The code base previously used home-grown linked lists, but I think it would be nice to use an existing library rather than revert the code.

Skills needed

C Programming

Difficulty level

Medium

Project size

Medium (175 hours)

Mentors

Mark Wong

Expected outcomes

pg_top functions with ported tree.h code for at least Linux

References

pg_top git repository
tree.h, a license compatible red-black tree implementation from OpenBSD source
pg_top presentation

Performance Farm: BuildBot test result data transformation

Project Description

Buildbot is a continuous integration framework being used to proof the next generation of the PostgreSQL Performance Farm project. This project is to extract data from BuildBot's database of test results, the PostgreSQL git repository, and the data saved on the Buildbot Worker to transform it, and load it into a new reporting database so that is it easier to query results.

The contributor is not expected to be familiar with Buildbot's database schema prior to starting, or the details of the various tests that are being run. An introduction to some of those details will be part of the start of the project.

Skills needed

SQL
POSIX Shell Scripting
git

Difficulty level

Easy

Project size

Medium (175 hours)

Mentors

Mark Wong

Expected outcomes

Schema design of new reporting database
A set of SQL and shell scripts that can be used in the Performance Farm project

References

git
PostgreSQL
Example script building a static report for DBT-2 test results

Performance Farm: Web user interface for navigating results

Project Description

Develop a front end in interface in java script for navigating test data.

The Performance Farm prototype runs various tests against all supported branches of PostgreSQL, but the prototype doesn't really have a good interface.

Here is an example of a static way data is visualized for the results of a single system. The primary way test results are views are for on specific test (e.g. DBT-2), only on one system at a time (e.g. vanillaleaf), and at one specific scale. Then metrics from multiple git branches (e.g. HEAD, REL_17_STABLE, REL_16_STABLE, etc) may all be viewed at the same time, for any code change (i.e. commit) in the PostgreSQL repository.

Here is an excerpt of raw data from one specific test for one specific system:

   branch,revision,scale,ctime,metric,complete_at
   REL_13_STABLE,3850fcca69b5db0694ceb5d1134699dc247f201e,100,1708386677,564578.0,1728363218
   REL_13_STABLE,9061fd23c28faebcb29bdfb262975639715975c0,100,1708719713,557362.69,1728362912
   REL_13_STABLE,43cca9de9a0adf3fb47aaa6310cc0022a78eee8a,100,1708895707,570032.0,1728362591

An ideal interface will let the user:

select the branches to display
adjust the date range on the fly
mouse over any data point to see

 * commit hash
 * commit description

Skills needed

javascript

Difficulty level

Medium

Project size

Medium (175 hours)

Mentors

Mark Wong
Rajiv Harlalka

Expected outcomes

Working implementation of a java script interface

pgexporter: Extension support

Project Description

pgexporter [1] is a Prometheus [5] exporter for PostgreSQL [4]. This project looks to enhance its functionality with support for PostgreSQL extensions. There needs to be a general framework such that it is easy to add extensions or use different versions. Top extensions like pg_stat_statements should be supported by the distribution.

Skills needed

C
PostgreSQL

Difficulty level

Medium

Project size

Long (350 hours)

Mentors

Saurav Pal <palsaurav (dot) 2020 (at) gmail (dot) com>
Jesper Pedersen <jesper (dot) pedersen (at) comcast (dot) net>

Expected outcomes

Top PostgreSQL extensions working out-of-the-box in pgexporter releases

References

[1] https://github.com/pgexporter/pgexporter

[2] https://github.com/pgexporter/pgexporter_ext

[3] https://pgexporter.github.io/

[4] https://www.postgresql.org/

[5] https://prometheus.io/

pgmoneta: WAL record filtering

Project Description

Apply custom filtering rules to WAL records before generating new files.

Filter based on:

Transaction type (INSERT, UPDATE, DELETE, etc.).
Schema or table names to allow selective replication.
Custom conditions for targeted processing.

Expand with Cross-Version WAL Streaming where these files can be streamed to other versions of PostgreSQL based backups.

Skills needed

C
PostgreSQL

Difficulty level

Hard

Project size

Long (350 hours)

Mentors

Shahryar Soltanpour <shahryar (dot) soltanpour (at) gmail (dot) com>
Jesper Pedersen <jesper (dot) pedersen (at) comcast (dot) net>

Expected outcomes

Implement filtering mechanism that can generate PostgreSQL WAL files for different versions.

References

[1] https://github.com/pgmoneta/pgmoneta

[2] https://pgmoneta.github.io/

[3] https://www.postgresql.org/

pgmoneta: Incremental backup for PostgreSQL 13-16

Project Description

Implement functionality in pgmoneta and pgmoneta_ext such that an incremental backup can be taken from PostgreSQL 13 to 16.

This work can build upon the work done in the incremental backup support for PostgreSQL 17 inside of pgmoneta.

Skills needed

C
PostgreSQL

Difficulty level

Hard

Project size

Long (350 hours)

Mentors

Haoran Zhang <andrewzhr9911 (at) gmail (dot) com>
Jesper Pedersen <jesper (dot) pedersen (at) comcast (dot) net>

Expected outcomes

Incremental backup working inside pgmoneta for PostgreSQL 13 - 16

References

[1] https://github.com/pgmoneta/pgmoneta

[2] https://github.com/pgmoneta/pgmoneta_ext

[3] https://pgmoneta.github.io/

[4] https://www.postgresql.org/

pgmoneta: Clustering support

Project Description

Implement functionality in pgmoneta and pgmoneta_ext such that pgmoneta can operate in a clustered way meaning f.ex. the result of a backup operation is replicated to another pgmoneta node.

It should be possible to control which node replicates to which node(s).