Better telemetry could improve observability and AI decisions

Organizations are reconsidering how they collect and manage telemetry as technology environments grow more complex. In a TechRadar Pro Perspectives article, Mala Pillutla argues that collecting more data does not necessarily make problems easier to diagnose. Fragmented signals can slow investigations, while storing and processing information that adds little insight can increase costs.

The article recommends improving telemetry earlier in its lifecycle. Teams can filter out information they do not need, add context to useful signals and route relevant telemetry before it reaches downstream systems. This can reduce the volume teams handle and make the resulting information easier to use. OpenTelemetry standards can help teams gather signals consistently across tools and vendors, including in distributed environments.

The quality of telemetry also matters for AI systems used in software delivery and operations. The article says incomplete or cluttered data can produce poor insights, making automated decisions harder to trust and explain. Consistent, contextual information can help teams understand why a system made a decision and support confidence among technical teams and senior leaders.

With more relevant signals, teams may spend less time investigating noisy data, identify issues sooner and automate with greater confidence. Pillutla’s central argument is that organizations can strengthen observability by making better choices about telemetry at the point where it is collected, rather than simply collecting more.