Prequel: Software Errors Be Gone  ## Metadata - Author: [[The New Stack Podcast]] - Full Title: Prequel: Software Errors Be Gone - Category: #podcasts - URL: https://share.snipd.com/episode/c402f866-9df3-4e38-aa15-bd48c6884ae6 ## Document Notes ## Highlights - **Founders' Bug Hunting Background** - Lyndon Brown shares that after more than a decade at NSA hunting elusive bugs, he and Tony Meehan started Prequel. - They faced the challenge of balancing rapid development with controlling application failures, common to many engineering teams. Transcript: Alex Williams Awesome. Lyndon Brown So as you said it, we're the founders of Prequel, the first community-driven platform for discovering bugs, misconfiguration, and developer anti-patterns. We started Prequel after a decade plus at the National Security Agency, really hunting for elusive bugs, and then leading engineering teams at Elastic and companies like Mandiant, Acquired by Google. But much like other engineering teams, we found it difficult to really balance the speed of development with control over application failure. So we believe every team should be able to look for both frequent and obscure reliability problems with a lot less effort. ([Time 0:01:05](https://share.snipd.com/snip/0b601c52-bd2d-4f6f-883a-8192bfe65fce)) - **Insight on Common Reliability Enumerations** - Common Reliability Enumerations (CREs) provide a deterministic, declarative, and expressive way to find software problems. - CREs aim to help teams discover and communicate recurring production issues more effectively than traditional observability tools. Transcript: Lyndon Brown And that's where common reliability enumerations, CREs, come into play. So we're working to give teams a deterministic, declarative, expressive way of finding problems. ([Time 0:01:43](https://share.snipd.com/snip/00390c0b-c714-40d1-a154-5b66aefc3f37)) - **How to Use Prequel's preq Tool** - Use Prequel's open-source tool, preq, to detect problems in real time by analyzing logs with continuously updated CREs. - Run preq in your environment via kubectl plugin or standalone for multi-platform support and actionable problem reports. Transcript: Tony Meehan Let's do it. So real quick demo here. There are a couple of different ways that you can run and test with CREs. The first is you can use a kubectl plugin. Install that right on your command line here, the kubectl crew install preq, and that'll install preq right away. It works on Windows, Mac, Linux, all the major architectures and operating systems. You can also run it in standalone. So if you go to docs.prequel you can download it there directly from GitHub. And so let's show you what it looks like to detect the problem in Postgres. I'll take a look at my logs here on my server for Postgres and pipe that in as standard input into preq. And it's like grep on steroids. And so what we're doing here is we're looking through this input, this log here for problems. And we've detected CRE 2025-47, which is a critical issue. We can take a look at that report. And what we're seeing is a summary of the problem, its description and impact, the mitigation that the community and the problem detection community has put together for this, as well As the specific matches for the problem. You can kind of think of a CRE as a description of the problem and how to mitigate it. And the really cool thing about Preq is that you can run this with continuous updates, stay up to date on the latest CREs that are coming from the community and the Prequel Reliability Research Team. So that's it. Very quick demo, but that's how it works. ([Time 0:01:57](https://share.snipd.com/snip/476b100d-14e7-4d4a-8cab-7dcc927eb39a)) - **Prequel Improves Reliability Detection** - Prequel offers proactive and more complete reliability problem coverage via automatic CRE updates from the community and research team. - It reduces noise by combining weak signals into high-fidelity findings, easing engineers' burden to focus on real issues. Transcript: Lyndon Brown And we already have some users deploying it in their clusters, and they're seeing a number of benefits. The first is just more proactive, more complete coverage of reliability problems through automatic updates, new CREs coming in from the community, also coming in from Prequel's Research team. Also, less noise. One of the biggest complaints we hear from observability users is, hey, how do I even know what I need to pay attention to? And part of that is just the inability of those platforms to really combine multiple weak signals into high fidelity findings, and also the ability to reduce red herrings before a human Has to do anything. So we're really excited about the power of CREs and Preq being able to enable every engineering team to really unlock the power. And the last thing, as I mentioned, completely open source, Apache 2. So use it how you want to use it and contribute if you want to contribute. ([Time 0:03:45](https://share.snipd.com/snip/c0e40349-ce19-4eaa-b914-3da7f18af562))