Making progress, one check at a time

Since I publicly announced pgFirstAid in November 2025, the project is somehow at 477 (as of 2026-01-06) stars on GitHub! Honestly, I was absolutely shocked to see how quickly things escalated. I remember the anxiety of submitting the first public acknowledgement on LinkedIn and Mastodon. I figured a few people would like it, maybe some of my colleagues would give thumbs up, and life would move along. Yet, one thing led to another, a friend messaged me about someone posting the tool on Hacker News, and the next thing I knew the project was mentioned in the TLDR Newsletter.

So thank you all for sharing the project and taking this farther than I thought it would ever get, especially in the first few days after it was released! We even had some folks send in some issues and PRs that were eventually merged into the main branch! Hell yeah Open-Source!

Milestones 2 and 3

Over the next few weeks, I got to work on adding new health checks, the option to use pgFirstAid as a view, and reworked some of the existing code to align with the format that was taking shape as the function was built. Instead of listing all of the new health checks that were added, I encourage you to look at the list of health checks within the README. Something that is bugging me, currently, is the lack of “low” priority checks that are present. I will be focusing on adding some more of those over time, but I want to make sure the tool can provide (in its earlier stages) areas where users can take immediate action.

When looking back at these releases, I am happy to see progress and hope to see more contributions from members of the community (via PRs or feedback on current or future checks).

SCALE 23x (Southern California Linux Expo) Talk(s)

Last year I attended SCALE (22x) for the first time and immediately knew I wanted to submit a talk for the next year (2026). If you have not been to SCALE before, you should check out previous talks on their youtube channel or their website. One of the best parts about SCALE is that it includes multiple “paths” all wrapped into a single cost of admission. SCALE itself has “co-located” events that are focused on specific topics, like PostgreSQL @ SCALE, Nix/NixOs (Planet Nix), DevOps Days LA, Cloud Native Days LA, and so on. And if that was not enough to sell you, there is an expo that also takes place where vendors, open-source projects, and local businesses come in to talk to users! Yes, there is swag, but you also get to take advantage of one-on-one time with vendors that your company uses!

Anyways, I submitted two talks for this year’s SCALE and both were accepted! So I will be giving a talk on pgFirstAid (naturally) in the PostgreSQL @ SCALE event and a talk I called “Five Stages Of Grieving-Databases in Infrastructure as Code.” No spoilers here, so show up at SCALE and let’s talk databases!

Head here for a brief summary of the talks.

Future Work (+2 weeks)

Having served as a QA Test Lead in my earlier days in IT, it would be a shame if I did not bring up testing. Something I have done is spun up database servers running Postgres 15-18 and continue to use these as my test subjects when developing new checks or changes to pgFirstAid. Most of these databases are actually backends to tools we all use today. So, ironically, I think I have to start creating some tickets for these apps to add some indexes and primary keys…🤓

Regardless, I have some things I want to add to the workflow as new checks come in. The first thing I want to implement is utilizing Microsoft Azure, AWS RDS/Aurora, and GCP’s Cloud Database “free” tier offerings to run tests against the more common managed Postgres instances. This would mean deploying at least 4 databases per cloud provider and writing the modules using OpenTofu to manage the deployments. Thankfully, I have plenty of experience with AWS and IaC! However, I will be learning as I enter GCP and Azure (yes, I have a DP-900 certification under my belt but things have changed significantly since 2020). This will do for now, and it will be a manual process (deploying from a vm of sorts to avoid tying up my local machine). Eventually, it would be great to automate that—but one step at a time.

What would also be great is having a dataset that would trigger all of the health checks. This would give me the ability to restore the dataset, execute the function or view, and verify what I expect is present in the results. Some call it “golden data,” but the key is providing a 1:1 ratio to health-check:test.

Final Note

Something that may come off as selfish is I tried to reach out to the community to see if anyone was open to providing a logo for the project and a lot of people just pointed to chatty-jeeps (ChatGPT). Claude, and other AI Picture/Logo generation tools. Which is what I did, initially, to see if the tool could spit out anything resembling what was in my head that I would never be able to “draw” myself. Some of the logos were horrible (Claude, your image generation is terrible) while others provided some promise. Yet, in the spirit of Open-Source, I refuse to use any of these rendered images as a logo for pgFirstAid.

So, instead of asking for an artist to spend time on something they won’t get paid for, I would like to ask you, readers, if you could donate to the “fund” that would pay for this work?

The most direct way would be to donate by doing a one time sponsorship of me in GitHub (or recurring if you want I guess) or I will also get some sort of socially acceptable, totally not corrupt “goFundMe” fundraiser up to collect donations.

If you have an artist in mind, please reach out to [email protected]