The AI (artificial intelligence) landscape wouldn’t be where it is today without AI-as-a-service (AIaaS) providers like OpenAI, AWS, and Google Cloud. These companies have made running AI models as easy as clicking a button. As a result, more applications have been able to use AI services for data analysis, content generation, media production, and much more.

One of the biggest causes of outages and incidents is good old-fashioned human error. Despite all of our best intentions, we can still make mistakes, like forgetting to change defaults, making small typos, or leaving conflicting timeouts in the code. It’s why 27.8% of unplanned outages are caused by someone making a change to the environment. Fortunately, reliability testing can help you catch these errors before they cause outages.

If you've ever deployed a distributed system at scale, you know things break—often in ways you never expected. That’s where Chaos Engineering comes in. But running chaos experiments without robust observability is like debugging blindfolded. This guide will walk you through how observability empowers Chaos Engineering, ensuring that your experiments yield meaningful insights instead of just causing chaos for chaos’ sake.

When you think about “AI reliability,” what comes to mind? If you’re like most people, you’re probably thinking of generative AI model accuracy, like responses from ChatGPT, Stable Diffusion, and Sora. While this is certainly important, there’s an even more fundamental type of reliability: the reliability of the infrastructure that your AI models and applications are running on. AI infrastructure is complex, distributed, and automated, making it highly susceptible to failure.

Today, we’re excited to announce Gremlin Private Edition, a version of Gremlin that you can run entirely within your own private network.

Testing shouldn’t feel risky. While it might sound counterintuitive, certain types of testing can actually increase risks to your systems. Load testing, for example, is a great way to see how your systems behave under pressure, but it can also cause those same systems to fail if they aren’t equipped to handle the load. For some types of testing, this is necessary, as is the case with reliability testing and Chaos Engineering.

You’re well on your way to becoming more reliable. You’ve added your services, found and fixed some Detected Risks, and run your first set of reliability tests. However, some of your tests returned as “Failed.” Not to worry: this isn’t a reflection of you or your engineering skills but rather an opportunity to learn more about how your systems work and, more importantly, how to make them more resilient.

Cloud providers like AWS excel at creating reliable platforms for developers to build on. But while the platforms may be rock-solid, this doesn’t guarantee your applications will be too. It’s the provider’s job to offer stable infrastructure, but you’re still on the hook for making your workloads resilient, recoverable, and fault-tolerant. There’s only one problem: cloud platforms are essentially black boxes.

Reliability doesn’t happen by itself. Making a system reliable and resilient enough that your customers can count on it takes a combination of time, effort, and resources that could be used elsewhere, such as shipping new features. It’s also not optional. In an era where downtime costs an average of $14,056/min (or $843,360/hr), outages have a material impact on businesses. Unfortunately, most systems are sprawling and complex enough that even small amounts of downtime can add up quickly.

Reliability testing is ongoing work, and tracking that work can be difficult in large organizations. Engineers run one-off experiments, scheduled Scenarios run in the background, and, for more mature teams, CI/CD workflows fire off automated tests on demand. According to our own product metrics, teams run an average of 200 to 500 tests each day! With so much happening, it’s hard to keep track of everything going on in Gremlin—until now.