Grafana Labs has made Adaptive Profiles generally available in Grafana Cloud, completing its Adaptive Telemetry suite across metrics, logs, traces and profiles.
The product adjusts profiling detail and collection frequency based on workload behaviour. That lets engineering teams capture more detailed performance data during anomalies while keeping routine collection at a lower-cost baseline.
Grafana Labs is framing the launch around rising observability costs as companies collect growing volumes of operational data from software systems. The pressure is increasing as businesses adopt AI agents and applications built with large language models, which generate additional telemetry from model calls, tool use and automated workflows.
According to Grafana Labs' 2026 Observability Survey, 57% of organisations are already implementing LLM observability in some form, while 65% say cost is the main criterion when choosing observability tools.
Adaptive Telemetry is intended to address that pressure by analysing how telemetry data is used and recommending what should be kept, aggregated or dropped. With the addition of Adaptive Profiles, the approach now spans the four main observability signals used by software teams.
Steven Dungan, staff product manager at Grafana Labs, set out the company's position on the economics of observability.
"The fundamental problem with observability economics today is that cost scales with ingestion, not insight," Dungan said.
"Adaptive Telemetry inverts that model. Every signal - metrics, logs, traces and profiles - now has an intelligent layer that learns how data is used in practice and then optimizes automatically. With Adaptive Profiles reaching GA, we've closed the loop on the full stack. Teams get more signal, less noise and lower bills, and they don't have to sacrifice one for another."
Profiles added
Continuous profiling shows how applications consume CPU, memory and other resources in production systems. But broad deployment has often been limited by cost, particularly when profiling runs at high resolution across large infrastructure estates.
Adaptive Profiles addresses that by varying data collection automatically. Under normal conditions, it gathers profiling data at a lower level. When a performance problem or anomaly appears, it increases the resolution so engineers have more information to investigate.
Grafana Labs argues this makes wider use of profiling more financially practical for teams that have struggled to justify fleet-wide deployment.
Upland Software said cost control had been a key concern.
"Adaptive Profiles ensures that we can leverage Cloud Profiles without worrying about cost overruns," said Michael Beltz, vice president of cloud operations at Upland Software.
"The ability to see where code is slowing down, memory is being allocated, and where improvements are needed [with Cloud Profiles] is necessary for us to reduce infrastructure resources and improve the user experience."
Existing results
Adaptive Profiles joins three other tools in the suite that are already generally available: Adaptive Metrics, Adaptive Logs and Adaptive Traces. According to Grafana Labs, those products have delivered measurable reductions in data volumes and spending among Grafana Cloud users.
Adaptive Metrics is the most widely deployed of the four, the company said. It has removed 28.5 billion active series and delivered an average 35% reduction in metrics costs.
Mux was cited as one customer that cut its metrics volume by 60% and extended retention from 14 days to 13 months.
"Adaptive Metrics is an amazing feature. It not only saves us hundreds of thousands of dollars a year, but it's also a forcing function for us to look closely at our metrics to find additional opportunities for time series reduction and cardinality improvements," said Kyle Weaver, staff software engineer at Mux.
Adaptive Logs examines log usage to identify high-volume patterns that are rarely used and can be removed. Grafana Labs said the feature has eliminated 26 petabytes of log volume across its cloud platform.
TeleTracking, an early adopter of the tool, has seen a 50% reduction in log volumes, according to Grafana Labs.
"Adaptive Logs helps reduce noise, making it easier to spot valuable logs and ultimately saves us costs," said Andrew Qu, software engineer II at TeleTracking.
Adaptive Traces, which became generally available in late 2025, uses tail sampling to keep traces linked to errors, latency and other notable events while filtering lower-value, repetitive spans. Grafana Labs said this has reduced trace data volume by an average of 82% across Grafana Cloud.
Auditboard said the tool changed the trade-off between visibility and cost.
"Before Adaptive Traces, we had two bad options: send everything and blow our budget, or send so little we couldn't get meaningful insight," said Geoff Schultz, manager of infrastructure engineering at Auditboard.
"Now tracing is actually usable, we can dial sampling up or down as needed, keep costs in check and still give teams the visibility they need."
Across customers using multiple parts of the Adaptive Telemetry suite, Grafana Labs said total telemetry costs have fallen by 30% to 50% on average, with savings redirected into broader observability coverage.