Berlin, Germany
OllyGarden turbo charges mission to fix bad telemetry with $4m raise and launch of “Minimum Viable Instrumentation”
OllyGarden has raised $4m in VC funding and launched Minimum Viable Instrumentation (MVI), which lets Rose, its AI OpenTelemetry® Engineer, establish an instrumentation baseline where none exists yet.
OllyGarden, the platform helping fix bad telemetry to give observability engineers data they can rely on, has raised $4m in VC funding to take its capabilities to the next level.
The round was backed by leading VCs including Next Frontier Capital, Grand Ventures and ACTAI. Existing investors DIG Ventures doubled down on the telemetry pioneers, as did observability leaders Datadog, Grafana Labs, and Dash0 who made further strategic investments into the platform.
Launched in 2025, OllyGarden is working with companies across retail, financial services and enterprise software to end noisy, messy, and expensive telemetry systems which are churning out higher volumes of bad or unreliable data.
OllyGarden helps companies move from firefighting to future-proofing. The platform continuously evaluates telemetry quality, identifies issues, and then turns findings into source-level fixes. This unlocks useful, cost-effective operations across organisations by stopping bad telemetry data from reaching observability backends in the first place.
In practice, it means engineers and AI agents can make decisions based on good, clear data and stop bad information from seeping into wider operations.
Over the past year, OllyGarden has helped companies identify volume reduction opportunities across up to 85% of their logs.
In addition to the new capital, OllyGarden is today also announcing the launch of its latest product innovation: Minimum Viable Instrumentation (MVI). Until now, OllyGarden’s AI OpenTelemetry Engineer, Rose, has been able to review existing instrumentation and make suggestions to improve it. Now, with the launch of MVI, Rose can also detect where no existing instrumentation exists and provide a path to establishing an instrumentation baseline.
This innovation enables best practice instrumentation to exist in every corner of an organisation: in every team, every repository, and across every pull request. Crucially, it keeps things at a minimum; calibrating what’s needed to provide a foundation teams can actually operate on, rather than creating more noise for engineers to battle through.
“More instrumentation is not better instrumentation,” comments OllyGarden founder, Juraci Paixão Kröhling. “It’s about surfacing only the information and data that’s useful and actionable: no more, no less. The dawn of AI generated code has, unfortunately, created a mirror effect across telemetry. Observability engineers and their AI agents are awash with high volumes of bad data they don’t need, whilst critical information gets missed in the noise. And they’re paying through the nose to store and query it all.
“That’s why OllyGarden was founded - to weed out bad telemetry and give engineers a clear, simple route to insight that can scale across an organisation. And that’s also why we wanted to push our product capabilities even further with the launch of MVI; making it possible to drive results across existing instrumentation and pinpoint where instrumentation doesn’t yet exist but needs to. It’s OpenTelemetry expertise, applied consistently across every codebase you own.
“OpenTelemetry is the industry standard for software telemetry. As a team of OpenTelemetry experts, we know what best practice looks like. Our knowledge is encoded in our products and can be deployed at scale across large organisations even when they only have a small number of observability engineers.”
Minimum Viable Instrumentation launches today and is available to all existing enterprise customers, with a free trial available to new customers.
Investor quotes
Telemetry quality is a foundational enterprise problem that everyone has and nobody owns, and now AI has made it urgent. When the basics cannot be relied on, everything built above them breaks, and it breaks faster now because there is no human in the loop to catch it. Having backed Juraci and the team at OllyGarden's inception, we are excited to double down.
Every wave of AI has needed someone to make its input data trustworthy before the wave could deliver on its promise. Agentic operations are at exactly that inflection point. OllyGarden is building the trust layer for the data AI agents to run on, and Juraci is the rare founder who helped create the open telemetry standards and who can now build a highly scalable business on top of it.
Every company is rebuilding its operations stack on the assumption that agents will run it. Almost nobody is asking what those agents will be reading. Telemetry has quietly become the data layer AI operates from, and it is the least governed data in the enterprise. OllyGarden is building the quality layer for it, and we have not seen a team better placed to do that.
I spent years building telemetry pipelines, and the lesson is that you cannot filter your way out of bad instrumentation. It has to be fixed where it is written. OllyGarden is the first team to treat that as a product rather than as a code review, and they are doing it at the moment when agents, not people, became the primary consumers of the data.
About OllyGarden
OllyGarden helps teams find and fix bad telemetry before it reaches any observability backend, giving engineers and AI agents data they can rely on, from pull request to production. Its mission is to help every company confidently say it generates, collects and stores the telemetry it needs, without excess. Founded in 2025 by Juraci Paixão Kröhling, an OpenTelemetry Governance Committee member, OllyGarden is built by OpenTelemetry contributors who help shape the standard.
Learn more at ollygarden.com.