Sawmills.ai Releases 2025 State of Observability & Telemetry Report: Enterprises Spend Big on Observability, But on Average, Only 13% of Telemetry Data is Being Used
Sawmills' 2025 State of Observability & Telemetry Report surveyed DevOps and engineering leaders and found that only 13% of telemetry is ever used, while ballooning data volumes are driving up costs and muddying signal-to-noise. The findings highlight both the opportunities for optimization—and the growing role of AI—in fixing observability at scale.
San Mateo, CA, September 11, 2025 --(PR.com)-- Sawmills.ai, the first AI-powered telemetry management platform, today released its 2025 State of Observability and Telemetry Report, which found that mid-market and enterprise companies are ingesting record volumes of telemetry data, yet most of it goes unused. The result is not only ballooning observability costs, but also a degraded signal-to-noise ratio that makes it harder for engineers to detect, investigate, and resolve critical issues quickly.
The survey of senior DevOps and engineering leaders in the US and EU shows that organizations are struggling to separate signal from noise in their monitoring pipelines. On average, just 13% of collected telemetry is actively used, while 84% of companies admit they use less than a quarter of what they ingest. Despite this, the typical enterprise spends at least a million annually on observability, with many topping that figure.
Commentary on the Findings
“The industry has been conditioned to send everything to their observability and SIEM solutions, but the economics no longer add up,” said Ronit Belson, CEO of Sawmills.ai. “Teams are drowning in logs, traces, and metrics that add no value, yet every gigabyte drives up the bill. And this flood of junk data actively hurts reliability by burying the signals engineers need to resolve incidents quickly.”
The survey highlights how tool sprawl compounds the problem: 86% of companies use two or more observability platforms, often blending commercial solutions like Datadog (38%) and New Relic (41%) with open-source tools such as Grafana, Loki, and Prometheus (48%). While intended to improve coverage, this mix creates fragmented pipelines and redundant ingestion, which contribute to increased costs and complexity.
The Rise of AI in Observability
The report also reveals a strong appetite for AI-driven approaches: 63% of companies are interested in building AI agents to manage and optimize telemetry, and 60% are exploring AI copilots for SRE teams to accelerate incident response.
“This is a tipping point,” Belson added. “Teams recognize that manual filtering and rule-writing can’t keep up. AI is uniquely suited to act as a guide that sends the right telemetry to their observability solution in real time, giving engineers clarity and CFOs relief from runaway costs.”
Key Takeaways from the Report
- Data waste is endemic: On average, only 13% of collected telemetry data is being used
- Costs are ballooning: Enterprises spend nearly $1M annually on observability.
- Tool sprawl fuels inefficiency: Most companies juggle multiple platforms, driving redundancy.
- Signal-to-noise is a priority: 48% of leaders focused on improving data quality and of those 61% are not satisfied with their observability provider
About the Survey
The survey was conducted in June 2025 by Global Surveyz Research with responses from DevOps, SRE, Platform, and Cloud engineering leaders at organizations with 500+ employees across the US, UK, Germany, France, and Italy.
About Sawmills
Sawmills is redefining telemetry for the age of AI. As the first AI-powered telemetry management platform, Sawmills helps DevOps teams cut observability costs by up to 80% while improving data quality and system reliability. With real-time insights, one-click fixes, and automated policies, Sawmills empowers teams to take back control of their telemetry pipelines.
Download the full report here:
https://www.sawmills.ai/observability-report-2025
Media Contact:
winston@sawmills.ai
The survey of senior DevOps and engineering leaders in the US and EU shows that organizations are struggling to separate signal from noise in their monitoring pipelines. On average, just 13% of collected telemetry is actively used, while 84% of companies admit they use less than a quarter of what they ingest. Despite this, the typical enterprise spends at least a million annually on observability, with many topping that figure.
Commentary on the Findings
“The industry has been conditioned to send everything to their observability and SIEM solutions, but the economics no longer add up,” said Ronit Belson, CEO of Sawmills.ai. “Teams are drowning in logs, traces, and metrics that add no value, yet every gigabyte drives up the bill. And this flood of junk data actively hurts reliability by burying the signals engineers need to resolve incidents quickly.”
The survey highlights how tool sprawl compounds the problem: 86% of companies use two or more observability platforms, often blending commercial solutions like Datadog (38%) and New Relic (41%) with open-source tools such as Grafana, Loki, and Prometheus (48%). While intended to improve coverage, this mix creates fragmented pipelines and redundant ingestion, which contribute to increased costs and complexity.
The Rise of AI in Observability
The report also reveals a strong appetite for AI-driven approaches: 63% of companies are interested in building AI agents to manage and optimize telemetry, and 60% are exploring AI copilots for SRE teams to accelerate incident response.
“This is a tipping point,” Belson added. “Teams recognize that manual filtering and rule-writing can’t keep up. AI is uniquely suited to act as a guide that sends the right telemetry to their observability solution in real time, giving engineers clarity and CFOs relief from runaway costs.”
Key Takeaways from the Report
- Data waste is endemic: On average, only 13% of collected telemetry data is being used
- Costs are ballooning: Enterprises spend nearly $1M annually on observability.
- Tool sprawl fuels inefficiency: Most companies juggle multiple platforms, driving redundancy.
- Signal-to-noise is a priority: 48% of leaders focused on improving data quality and of those 61% are not satisfied with their observability provider
About the Survey
The survey was conducted in June 2025 by Global Surveyz Research with responses from DevOps, SRE, Platform, and Cloud engineering leaders at organizations with 500+ employees across the US, UK, Germany, France, and Italy.
About Sawmills
Sawmills is redefining telemetry for the age of AI. As the first AI-powered telemetry management platform, Sawmills helps DevOps teams cut observability costs by up to 80% while improving data quality and system reliability. With real-time insights, one-click fixes, and automated policies, Sawmills empowers teams to take back control of their telemetry pipelines.
Download the full report here:
https://www.sawmills.ai/observability-report-2025
Media Contact:
winston@sawmills.ai
Contact
Sawmills
Winston Bowden
919-219-9610
www.sawmills.ai
Winston Bowden
919-219-9610
www.sawmills.ai
Contact
Categories