Chaos experiments have revealed weak points and now provide controlled cost-saving testsThe best features of Gremlin Reliability Management Platform are the safe failure injection, which is crucial as we can simulate the failures in a manner that we know these are just dumping tests and not the actual issues. Whether it is the CPU spike or the memory exhaustion, or the network latency, or the server shutdown, server shutdown is one of the most favorite features that I have in Gremlin Reliability Management Platform. The controlled blast radius is another standout feature. The controlled blast radius feature has helped my team in that we actually wanted to target only one specific container, our Docker containers that we deployed. It helped us to conduct tests in a very specific, isolated manner instead of launching a larger test or focusing on hundreds of servers at a time, resulting in very limited impact. Since ours is a very small team, we do not want to impact other servers. This controlled blast radius helped us to only focus on our servers and not impact any other team. Gremlin Reliability Management Platform has positively impacted my organization because before Gremlin Reliability Management Platform, we did not even know how to conduct these chaos engineering tests. We heard about it, but we had no idea of how to do something of that nature. If there are ten servers, ten systems in our architecture and if suddenly something goes down, nobody knew what would happen next. We did not even know how to simulate these types of tests. This lack of confidence has been mitigated by using Gremlin Reliability Management Platform. Now we can confidently test and see which system is the most critical. If this goes down, what happens? How much business valuation are we going to impact? How much loss are we going to incur? All of this is now clearly visible and transparent. Since using Gremlin Reliability Management Platform, we were able to reduce the incidents by six percent after conducting our limited experiments. We were also able to increase the uptime from ninety-eight to ninety-nine, which represents a one percent increase in uptime.
