Gremlin, the world’s first hosted Chaos Engineering service created by engineers formerly at Amazon and Netflix, announced ‘Gremlin Free’ which makes it easy for DevOps teams to get started with Chaos Engineering.
Gremlin Free includes the ability to randomly shut down servers, like the open source Chaos Monkey tool from 2011, but can also target specific hosts and simulate CPU spikes for more focused experiments. It also comes with a simple user-interface (UI), a halt button to safely rollback attacks, and industry-leading security not available with open source solutions.
Downtime is expensive: Gartner cites average per-company figures at $5,600 per minute (roughly $300,000 per hour), and for top eCommerce websites that figure can be millions per hour. The Gremlin team, made up of engineers and on-call leaders from Amazon, Netflix, Google and Dropbox, have developed tools that enable teams to better understand their systems and identify weaknesses before they cause outages and impact customers.
“Similar to a flu shot, the idea is to purposefully inject a controlled bit of harm in order to build up an immunity,” said Lorne Kligerman, Director of Product at Gremlin. “It’s still a new concept to most engineering teams, so we wanted to offer a free version of our software that helps them become more familiar with Chaos Engineering — both from a tooling and culture perspective.”
“I love Chaos Engineering and Gremlin’s service is the best there is,” said Charity Majors, CEO and Co-Founder of Honeycomb. “Injecting controlled failure into your systems to examine how they behave aligns perfectly with the goals of observability.”
Before launching Gremlin, CEO and Co-Founder Kolton Andrus worked at Netflix and built their second generation of fault-injection tooling (F.I.T.) to go beyond the limitations of Chaos Monkey. To date, Gremlin has raised $26.8Million in funding from Redpoint Ventures, Index Ventures, and Amplify Partners.
Existing customers include Expedia, Qualtrics, Twilio, Under Armour, and Walmart. Gremlin Pro offers a dozen different attack modes to better simulate real-life scenarios, as well as Application-Level-Fault-Injection (ALFI) for more granular attacks that work in serverless environments.