The Observability of Observability – O’Reilly

Regardless of the promise of AIOps, the dream of absolutely automated, self-healing IT environments stays elusive. Generative AI instruments often is the resolution that lastly abstracts away sufficient of the workload to get there. Nevertheless, at present’s actuality is much extra advanced. Web efficiency monitoring agency Catchpoint’s latest SRE Report 2025 discovered that for the primary time ever, and regardless of—or maybe due to—the rising reliance on AI instruments, “the burden of operational duties has grown.”

True, AI can easy out thorny workflows, however doing so might have surprising knock-on results. For instance, your system might use discovered patterns to routinely suppress alerts, however this might trigger your groups to overlook novel occasions completely. And AI gained’t magically repair what’s outdated or damaged: After implementing an AI resolution, “points typically stay as a result of change occurs over time, not instantly,” Catchpoint’s Mehdi Daoudi defined to IT Brew. That’s partly as a result of “making correlations between [the] totally different knowledge sorts residing in numerous knowledge shops is error-prone and inefficient” even with the help of AI-powered instruments, write Charity Majors, Liz Fong-Jones, and George Miranda of their forthcoming version of Observability Engineering. And that is earlier than bearing in mind the broader fear that overreliance on AI methods and AI brokers will result in the widespread erosion of human experience.

It’s secure to say AIOps is a double-edged sword, slicing via advanced processes with ease whereas introducing new types of hidden complexity on the backswing. As with generative AI as a complete, the utility of an answer most frequently hinges on its reliability. With out perception into how AI instruments are arriving on the choices they make, you’ll be able to’t make certain these choices are reliable. Michelle Bonat, chief AI officer at AI Squared, calls this “the paradox of AI observability.” In brief, as we delegate observability to clever methods, we scale back our skill to know their actions—or our monitoring methods. What occurs then after they fail, turn into unreliable, or misread knowledge? That’s why we’d like observability of our observability.

Why “Observability of Observability” Issues

IT ops groups are placing extra of their belief in automated alerts, AI-driven root trigger evaluation, and predictive insights, however this confidence is constructed on shaky floor. There are already issues about how efficient present AI benchmarks are at assessing fashions, and benchmarks for AI brokers are “considerably extra advanced” (and subsequently much less dependable). And observability presents its personal task-specific problems:

The integrity of your knowledge and knowledge pipeline: If the information sources feeding your observability platform are defective (e.g., dropped logs, misconfigured brokers, excessive cardinality points from new providers) or if knowledge transformation pipelines inside the observability stack introduce errors or latency, you’re in hassle from the beginning. You’ll be able to’t tackle the issues you don’t see.

Mannequin drift and bias: AI fashions are inclined to degrade or “drift” over time, due to adjustments in system conduct or knowledge, new utility variations, or rising discrepancies between proxy metrics and precise outcomes. And bias is a frequent downside for generative AI fashions. That is significantly vexing for observability methods, the place correctly diagnosing points calls for correct evaluation. You’ll be able to’t belief the output from an AI mannequin that develops biases or misinterprets alerts from the information, however as a result of LLM-in-observability platforms can’t typically clarify how they attain their conclusions, these points might be onerous to identify with out metaobservability.

Platform well being and efficiency: Observability platforms are advanced distributed methods—they’ve outages, efficiency degradation, and useful resource rivalry like some other. Protecting your main supply of fact wholesome and performing reliably is essential. However how will you realize your monitoring instruments are working correctly with out observability into the observability layer itself?

Your Observability Stack Is a Vital System. Deal with It That Means.

The answer is easy sufficient: Apply the identical monitoring rules to your observability instruments as you do to your manufacturing purposes. In fact, the satan’s within the particulars.

Metrics, logs, and traces: Telemetry knowledge offers you perception into your system’s well being and exercise. You have to be monitoring platform latency, knowledge ingestion charges, question efficiency, and API error charges in addition to AI-focused metrics like useful resource utilization of brokers and collectors, time to first token, intertoken latency, and tokens per second if relevant. Gathering logs out of your observability elements will assist you to perceive their inner conduct. And you’ll establish bottlenecks by tracing requests via your observability pipeline.

Knowledge validation and high quality checks: Standardizing observability knowledge assortment and consolidating your knowledge streams offers stakeholders a unified view of system well being—important for understanding and trusting AI-driven choices. OpenTelemetry is a significantly good platform for observability, because it provides portability in your knowledge, obviates vendor lock-in, and promotes constant instrumentation throughout various providers; it additionally permits higher explainability by linking telemetry to choice origin factors. However remember to additionally implement automated checks on the standard and completeness of information flowing into your observability instruments (variety of distinctive service names, anticipated metric cardinalities, timestamp drift, and many others.) in addition to alerts for anomalies in knowledge assortment itself (e.g., sudden drop in log quantity from a service). Like AI fashions themselves, your configuration will drift over time (an issue lower than one-third of organizations are proactively monitoring for). As Firefly’s Ido Neeman notes in The New Stack, “Partial IaC [Infrastructure as Code] adoption blended with systematic ClickOps mainly ensures configuration divergence.”

Mannequin monitoring and explainability: Honeycomb’s Austin Parker argues that the velocity at which LLM-based observability instruments can present evaluation is the true sport changer, although “they may be mistaken a dozen occasions earlier than they get it proper.” (He’ll be discussing how observability can match the tempo of AI in additional element at O’Reilly’s upcoming Infrastructure & Ops Superstream.) That velocity is an asset—however accuracy can’t be assumed. View outcomes with skepticism. Don’t simply belief the AI’s output; cross-reference it with easier alerts, and don’t low cost human instinct. Higher but, demand insights into mannequin conduct and efficiency, similar to accuracy, false positives/negatives, and have significance.¹ It’s what Frost Financial institution CISO Eddie Contreras calls “high quality assurance at scale.” With out this, your AI observability system will probably be opaque—and also you gained’t know when it’s main you astray.

The Evolving Position of the Engineer

AI is including new layers of complexity and criticality to IT ops, however that doesn’t diminish the software program engineer’s function. Ben Lorica contends that the “‘boring’ fact about profitable AI” is that “the winners. . .will probably be outlined not simply by the brilliance of their fashions, however by the quiet effectivity and resilience of the infrastructure that powers them.” Contemplating this “fact” from one other angle, CISO Collection host David Spark asks, “Are we creating an AI-on-AI arms race when what we actually want is fundamental engineering self-discipline, logging, boundaries, and human-readable perception?”

Good engineering practices will at all times outperform “utilizing AI to resolve your AI issues.” As Yevgeniy Brikman factors out in Fundamentals of DevOps and Software program Supply, “Crucial priorities are sometimes safety, reliability, repeatability, and resiliency. Sadly, these are exactly GenAI’s weak areas.” That’s why the quiet reliability Lorica and Spark champion requires steady, intentional oversight—even of instruments that declare to automate oversight itself.² Engineers at the moment are the arbiters of belief and reliability, and the long run belongs to those that can observe not simply the appliance but additionally the instruments we’ve entrusted to observe it.

Footnotes

1. For an in depth take a look at what’s required, see Chip Huyen’s chapter on evaluating AI methods in AI Engineering and Abi Aryan’s overview of monitoring, privateness, and safety in LLMOps. Aryan will even share methods for observability at every stage of the LLM pipeline at O’Reilly’s upcoming Infrastructure & Ops Superstream.
2. Simply the place people belong within the loop is an open query: Honeycomb SRE Fred Hebert has shared a helpful listing of inquiries to assist you to determine it out in your particular circumstances.