Israeli startup DataAgent Ltd. formally launched today with $10 million in pre-seed funding and a platform that autonomously repairs production faults inside a customer's own Kubernetes clusters. The company, founded in January by Ishay Yaari and Nati Shalom, aims to cut observability costs and speed up fault resolution. The funding will be used to accelerate customer adoption in North America.
Both founders previously worked at Cloudify Platform Ltd., the open-source cloud orchestration company that Dell Technologies acquired in 2023 for a reported $100 million. DataAgent's approach flips the traditional observability model. Conventional tools detect a fault and hand it to an engineer. That model requires a paid second copy of logs and metrics shipped to the vendor's cloud, which grows more expensive as systems scale.
The cost problem in observability
Observability spending has become a major line item for enterprises. A Grafana Labs 2025 observability survey found that spending averages 17% of compute infrastructure spending, with 10% as the most common figure. DataAgent claims customers can cut up to 90% of their observability bill by running the platform alongside existing tools.
Yaari, now co-founder and CEO of DataAgent, put the problem bluntly. "In 10 years, no one has ever heard an engineering leader say their organization's observability costs went down and that is no accident," he said. He tied the issue directly to vendor incentives. "When a vendor's revenue is your data ingest, it cannot cut your bill without cutting its own."
Shalom, co-founder and CTO, framed the industry's trajectory in stark terms. He said the sector has spent years "building better ways to watch production and charging more for it every year," and described autonomy as an architectural shift, not a feature layered onto observability.
How the platform works
DataAgent's software runs inside cloud-native control planes and sits above a customer's existing monitoring stack. Its agents read live system state, topology, and configuration drift where the data sits. The agents determine the cause of an incident and apply pre-approved fixes via guardrails defined by the customer. Deeper root-cause analysis runs offline after the service is restored.
Data goes outward only when a fault needs deeper inspection. The agent is open source and can be run standalone at no cost. A paid SaaS tier handles fleet management and orchestration. The design answers enterprise resistance to running closed software in production.
During onboarding, a discovery phase shows which failure classes the software can fix. Faults not proven correct are routed to a human. Approved fixes deploy through an existing change management process or a single CLI instruction. DataAgent claims its ordering, fix first and analyze later, shortens mean time to resolution.
Trust through staged autonomy
The staged approach builds trust gradually. The discovery phase sets expectations about what the software can handle. Human routing ensures that unproven fixes never reach production without oversight. The open-source agent addresses the security concern of running proprietary code inside sensitive clusters.
Shalom described the shift as fundamental. Autonomy, he argued, is not an add-on to existing observability tools. It changes where analysis happens and who acts on it. The platform is designed to work with existing monitoring stacks, not replace them.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
Investor backing and market context
MizMaa Ventures Ltd. and Alicorn Venture Partners co-led the pre-seed round. MizMaa also backed Tenet Security Inc., a cybersecurity startup that launched in June with $6 million in seed funding. Tenet focuses on policing rogue AI agents, a different slice of the same autonomous-systems wave.
The funding round arrives as enterprises grapple with rising infrastructure costs and a growing number of production failures. DataAgent's pitch is simple: keep the data where it lives, fix what can be fixed automatically, and send only the hard cases to humans.
The startup's model challenges the economics of traditional observability. Vendors that charge per byte of ingested data face pressure as customers look for ways to reduce spend. AI copilots added to traditional observability tools have not changed the cost outcome, according to the analysis in the original report.
DataAgent's agents run where the data sits, which means less data leaves the cluster. The paid tier adds fleet management and orchestration for organizations running the agent across many clusters. The open-source option lets smaller teams test the approach without upfront cost.
The company's claim of up to 90% observability cost reduction is significant, though it depends on customers running the platform alongside existing tools. The discovery phase during onboarding helps determine which failure classes the software can fix, which sets realistic expectations.
Launch details and next steps
DataAgent formally launched on September 01, 2026. The company started in January, giving the founders roughly eight months to build the platform and secure funding. The $10 million pre-seed round will support customer adoption efforts in North America.
The founders bring experience from Cloudify, where they built orchestration tools for cloud environments. Dell's reported $100 million acquisition of Cloudify in 2023 validated that market. Now they are applying similar automation instincts to the observability layer.
The article was updated at 05:00 EDT on September 01, 2026. The author is Duncan Riley. Image credit goes to DataAgent.
SiliconANGLE, the publisher, reaches 15 million+ viewers of theCUBE videos and counts 11,400+ theCUBE alumni among technology and business leaders. SiliconANGLE Media was founded by John Furrier and Dave Vellante. The company maintains an AWS marketplace portal for support. Latest stories include a Bank of England AI warning, OpenAI ad revenue, CrowdStrike Falcon IQ, the FTC Amazon lawsuit, Amazon Graviton5, and an Nvidia MediaTek investment.