For all their benefits, modern development practices (DevOps, microservices architectures, increased deployment frequency) increase problem diagnosis and resolution complexity. The typical approach to troubleshooting, leveraging centralized observability data, may not catch the root cause of most problems—a code or configuration change. Not quickly pinpointing the root cause of the problem prolongs the issue and may result in temporary fixes, repeat occurrences and longer downtime. The so-called “troubleshooting tax”—the amount of time engineers spend on
troubleshooting—remains hard to quantify and a challenge to minimize.