Over the last decade, significant investments have been made in the large-scale observability systems, accompanied by the widespread adoption of the discipline of Site Reliability Engineering (SRE). Yet, an over-reliance on observability alone has led us to a plateau, where we are witnessing diminishing returns in terms of overall reliability posture. This is evidenced by the persistent and prolonged app failures even at well-resourced companies that follow the best practices for observability.
Furthermore, the quest for reliability is forcing companies to spend ever more on observability, rivaling the costs of running the services they aim to monitor. Commercial SaaS solutions are even more expensive, as the unpredictable pricing models can quickly skyrocket the observability bill. The toll isn't only monetary; it extends to the burden shouldered by developers implementing observability and operators tasked with maintaining a scalable observability stack.