Allen Plummer

Choosing a distributed lock manager for your Go application? I tested 4 options in production (50K+ ops/day): **Performance**: Redis (~5ms) > PostgreSQL (~8ms) > etcd (~15ms) > Raft (~25ms) **Consistency**: etcd = PostgreSQL = Raft > Redis **Complexity**: Redis = PostgreSQL < etcd < Raft My choice by scenario: **Redis** → high load, eventual consistency acceptable **etcd** → K8s, leader election, consensus required **PostgreSQL** → already in stack, strong consistency critical **Custom Raft** → specific requirements, full control After 2 years in production: zero data corruption, 99.9% uptime. **Golden rule**: always set TTL. Full comparison with code: https://lnkd.in/d8vByGUa #golang #backend #architecture

137

7 Comments

At the last KCDC, I did a talk on "Modularizing the Monolith". Probably half the time was spent just on techniques for discovering and defining boundaries. People will still get boundaries terribly wrong with modular monoliths but the stakes are much lower. If you got your boundaries wrong with your microservice architecture, guess what! They'll still be wrong in your modular monolith. You'll still have a big ball of mud, but at least it's not a distributed monolith? Discovering and defining boundaries is hard. Distributed systems are harder. Doing both is harder-er.

158

6 Comments

I had to do a double take after reading the title for this article on "Pentagon declares war on open source", which would be a colossal mistake. But thankfully its over sensationalized and seems to just be referring to a single sentence from Katie Arrington's memo where she correctly acknowledges the risk unverifiable OSS poses to the DoD (screengrab below). OSS comes with its own set of challenges for everyone, but the DoD is a uniquely sensitive environment when it comes to the risks associated with running code written by foreign or (more commonly) anonymous contributors. OSS isn't going to go away, but thankfully there is a path forward to safely using it in critical systems, and I'm proud to be working on making that possible here at Chainguard. The solution is simple, but that doesn't make it easy. Build everything from source to reduce the supply chain risk dramatically by vertically integrating and eliminating attacks at the build system and package manager level. Audit the source code itself instead of relying maintainer identity and trust. Back to the basics. https://lnkd.in/ePP7gEC7

107

14 Comments

Had a great 8:30 AM chalk talk on building an effective test strategy for your serverless and event-driven applications. It’s never too early to test! Grateful to speak about the work Dan Fox Brian Krygsman and the whole serverless testing team did here at #reinvent25. As a reminder - Testing EDA is easy with the use of test harnesses to capture and evaluate the end state of your processed events. Sample code in the GitHub repository linked here. https://lnkd.in/eCAxnEyY

93

3 Comments

Pro tip for open source: There are a few OSS projects I've leveraged in the enterprise that were maintained by a single person. Lodash is one. Mountebank another. They always made me nervous but the way they communicated with users and the care they took about docs (popularity isn't enough for me) gave me enough confidence to use them. I was recently starting to evaluate another open source project that's been getting traction recently. I asked Claude to give some feedback on it. - Alpha release with an unstable API. Fair enough, something to test out. It's very popular, so that will improve - Single maintainer Not awesome, but not a deal killer - Docs and implementation do not match Hmmmm - 300+ open issues Hmmm Then I talked to someone who had ask the maintainer to do some consulting - Self aggrandizing - Hard to tell if he's really smart or a confident BSer (Elon Musk Syndrome) - Makes wild claims about accomplishments and people stealing ideas Yeah, I'm out. Open source is a mixed bag. You need to understand the risk profile of your supply chain. I'm looking for a core capability for an enterprise, not a toy, and this project is a pass. No, I won't say which one in public for obvious legal reasons.

25

11 Comments

One area I consistently think QA would benefit from is having better control over the test data used in the databases their suites run against. LLMs have made writing seed files SO SO SO much easier. Video in the comments. I spent some time learning how to do this with Convex for my Playwright suite, and it turned out to be a great hands-on experience. The video I put together is very much a learn-as-we-go session, full of trial, error, and real-time problem solving. In it, I explore how to create ephemeral (preview) environments with Convex and use the init.ts file to seed those environments with data for a Playwright test suite. I started with no understanding of how Convex handled this, and by the end I was able to spin up a preview environment and run a working seeding script. If you're interested in how seeded data can be used effectively in preview environments, how to make use of Convex’s init.ts file, or just want to see how this process works from a real-world, unpolished perspective, this video might be useful. I’ll be following up with another video showing how this setup can be integrated into a CI/CD pipeline for automated testing. #qa #testing

50

18 Comments