As KubeCon + CloudNativeCon North America 2025 prepares to descend on Atlanta this November, the cloud-native community faces a growing challenge: operational complexity and signal overload are outpacing human ability to make sense of Kubernetes environments. Enter ControlTheory, a startup making waves with its fresh approach to observability that flips the traditional “collect everything, store everything, query everything” model on its head. After announcing $5 million in seed funding from Silverton Partners at KubeCon Europe 2025 and watching its open-source Gonzo TUI tool gather over 2,000 GitHub stars in just two months, the company is returning to KubeCon as a Startup Sponsor with something significant to share.
At Booth 1570, ControlTheory CEO Bob Quillin and his team will launch the public preview of Dstl8, an AI-driven observability platform powered by M�bius continuous AI that promises to help organizations “see more, store less, and respond faster.” Built on OpenTelemetry standards, Dstl8 performs edge distillation�analyzing telemetry locally to extract only essential information�then uses agentic AI to correlate patterns across clusters and explain incidents in plain English. In this exclusive VMblog Q&A, Quillin discusses how continuous AI is reshaping the observability landscape, why edge intelligence matters for scaling Kubernetes, and what enterprises should be planning for as observability, FinOps, and security converge into unified control planes in 2026 and beyond.
VMblog: Can you give us your elevator pitch? What key message will attendees hear from you at KubeCon NA 2025, and what actionable insights will they take back to influence their management teams?
Bob Quillin: ControlTheory builds practical AI tools for developers & SREs that make monitoring and observability faster, clearer, and smarter so you can more efficiently find and fix problems. Gonzo, our popular open-source TUI, gives developers real-time log visibility in a terminal UI. Our enterprise platform Dstl8, powered by M�bius continuous AI, distills telemetry at the edge and uses agents to correlate patterns, detect issues, and explain incidents.
VMblog: As a sponsor of KubeCon + CloudNativeCon North America 2025, what sponsorship level have you chosen, and what strategic objectives drove this investment?
Quillin: We’re returning to KubeCon as a Startup Sponsor to deepen our connection with the CNCF community that inspires our products, especially focused on OpenTelemetry and Kubernetes. For us, the investment is simple: this is the community driving the real-world applications and infrastructure our tools are built for. Our goal is to engage directly with practitioners, showcase new open-source tooling and AI-powered observability, and connect with partners advancing the CNCF ecosystem around OpenTelemetry, Kubernetes, and AI infrastructure.
VMblog: Where can attendees find you at the event? What hands-on demos, interactive experiences, or booth activities have you designed to showcase your technology?
Quillin: You’ll find us at Booth 1570, where we’ll be running live demos of Gonzo and previewing Dstl8. Attendees can try Gonzo’s real-time log analysis terminal UI. Gonzo is an open-source tool already trusted by thousands of engineers, gathering over 2000 GitHub stars in its first two months. We’ll also preview Dstl8, our AI-driven observability distillation solution, with hands-on examples showing how Dstl8 can distill all your logs at the edge, correlate that data across your environment, and leverage those insights through agents that identify and resolve incidents and explain anomalies across your environment.
VMblog: How has your company’s KubeCon presence evolved since your first participation? What value keeps driving your continued investment in this community?
Quillin: At the last KubeCon Europe, held earlier in 2025 in London, ControlTheory announced it had raised $5 million in seed funding from Silverton Partners, an Austin-based venture capital firm. At the event we demonstrated our observability control plane built on top of OpenTelemetry standards including the standard OTel collector, OTLP, and OpAMP to enable better operational and adaptive control of observability systems.
In August 2025 we contributed Gonzo to the community as an open source log analysis TUI project. At KubeCon, we’ll also be demonstrating our newest product Dstl8, powered by M�bius AI. M�bius is a multi-layer continuous AI platform that drives Dstl8. M�bius has been built on top of our original OTel control plane, extended by our own distribution of the OTel collector and customer processors, and designed for high performance distillation at the edge to specifically drive, prompt, and enable our agentic AI analyses.
VMblog: Can you dive deeper into your company’s core technologies? What specific challenges do you solve for KubeCon attendees in their day-to-day operations?
Quillin: We focus on making observability practical and efficient. Gonzo helps developers and SREs instantly see what’s happening in their apps and systems, right from the terminal. Dstl8 takes it further: it continuously analyzes logs across clusters, distills signals from noise, and uses agentic AI to correlate related events. All needle, no haystack.
VMblog: In an increasingly saturated cloud-native marketplace, what distinguishes your solution in late 2025? What’s your unique value proposition?
Quillin: For the last 25 years, monitoring and observability have consisted of collecting ALL your data, storing it in a data lake, and applying business intelligence to analyze it through complex queries, indexing, and brittle dashboards. ControlTheory flips that model on its head. Instead of collecting all your telemetry, we distill out the important details at the edge, only sending the essential information packaged explicitly for AI prompts and agentic analysis and exploration (M�bius agents or other AI tools through MCP integration).
VMblog: With GenAI workloads and LLM deployments reshaping cloud-native architectures, how does your solution address these AI infrastructure demands?
Quillin: Dstl8 was designed for high volume, complex infrastructures, especially AI workloads running on Kubernetes. M�bius, the continuous intelligence inside Dstl8, a multi-layer AI that distills, correlates, and investigates in real time. It begins at the edge, reducing raw telemetry into structured summaries of sentiment, severity, and signal flow. Across clusters, it aggregates and correlates patterns and trends. And at the incident layer, M�bius is the autonomous agent that analyzes, explains, and learns from every signal. The result: always-on intelligence that turns observability into understanding.
VMblog: What sessions, announcements, or networking opportunities are you most excited about at KubeCon NA 2025?
Quillin: We are super excited about Observability Day on November 10 to catch up with the latest innovations and customer use cases for OpenTelemetry and AI monitoring and observability. KubeCon provides a pulse check on where the cloud-native world is heading and this year the conversation around MCPs, LLMOps, and continuous AI-powered observability will set the stage for the next 12 months. We’re especially interested in sessions and discussions that explore OpenTelemetry semantic standards, OpAMP, AI SRE workflows, and AI on Kubernetes use cases.
VMblog: Which KubeCon NA 2025 technical tracks or keynote themes align most closely with your company’s strategic direction?
Quillin: We love the fried peach pies and onion rings at the Varsity in Atlanta. We’ll see you there when we’re not at our booth 1570.
In addition to food, ControlTheory’s mission aligns tightly with the tracks focusing on AI + Cloud-Native Infrastructure, Observability and Performance, and Platform Engineering. Our platform, Dstl8 powered by M�bius, directly intersects these domains, applying multi-layer continuous AI to improve signal quality, correlation, and explainability across distributed systems. We’re part of the growing movement that’s pushing observability closer to edge intelligence, closed-loop automation, and platform-level simplicity, all themes that we expect to dominate this year’s KubeCon discussions.
VMblog: What’s your executive pitch for CTOs and CIOs? How do you demonstrate measurable business impact and ROI?
Quillin: Dstl8 helps organizations see more, store less, and respond faster. By applying AI at the edge, we dramatically reduce telemetry volume while improving detection accuracy and time-to-explanation and fix. You’ll see measurable business impact and ROI in four areas:
- Scaling Ops Without Headcount: Dstl8 automates understanding across environments, reducing MTTR by 50% while keeping ops teams lean. Like another SRE for your team.
- Maintenance Free Dashboards: No more stale panels or complex queries. Dstl8 always reflects your changes & gives instant, contextual answers, no dashboards to maintain, no training required.
- Find & Fix Issues Continuously: Most tools alert you after something breaks. Dstl8 spots and explains regressions the moment they happen – then leads developers to the fix.
- Make Logs Speak Human: Dstl8 explains what happened and why – no queries, no experts. Understand issues faster and keep everyone in the loop.
Dstl8 is not just an observability solution, but more of an operational intelligence layer that continuously feeds clean, contextual data back into your developer, SRE, DevOps, FinOps, and AIOps functions. For technical leaders, the ROI is clear: staff scalability, more insight, less infrastructure drag and toil, and measurable acceleration in incident response and MTTR.
VMblog: Can you outline a typical customer implementation journey? What critical pain points do you resolve, and what outcomes do clients achieve?
Quillin: Most teams start with Gonzo, our open-source TUI, to get immediate, real-time visibility into their logs directly in their terminal workflow. Once they hit scale, expand their team, and/or shift to production, they reach for Dstl8. Deployment begins at the Edge Distillation layer, where agents start summarizing telemetry locally. Then, the Correlation Layer connects signals across services, and Incident Layer agents deliver explainable summaries to users, integrating with existing tools like Slack or PagerDuty. Customers see value immediately, including:
- 40-60% reduction in log and metric volume without losing context.
- Faster incident understanding: AI-generated correlation summaries cut triage time from hours to minutes.
- Clearer communication between engineering and management, since incidents are explained in natural, narrative form.
VMblog: How does your technology fit within the broader CNCF ecosystem? What’s your role in the modern cloud-native infrastructure stack?
Quillin: Dstl8 is built on OpenTelemetry and our underlying AI platform M�bius aligns with CNCF’s open-control-plane vision. In addition to open sourcing Gonzo, where OTel is the canonical input, we contribute upstream and extend the ecosystem by introducing agentic intelligence as a control layer. Think of it as turning your observability data plane into an adaptive feedback system that’s open, composable, and continuously learning.
Gonzo integrations are a great example of ways to work within the products and platforms you already have. That includes live tailing AWS CloudWatch logs or Grafana Loki logs, to tailing Kubernetes logs with K9s and Stern, to analyzing Vercel logs.
VMblog: Are you unveiling any new products, features, or partnerships at KubeCon NA? What exclusive announcements can attendees expect?
Quillin: The public preview of Dstl8 launches at KubeCon. We’ll be demonstrating these Dstl8 capabilities:
- Continuous AI Analysis: Real-time, plain-English summaries of what changed, what broke, and why across your environment. No complex queries are required. Dstl8 continuously analyzes logs and distills only what you need.
- Root Cause Explanations: Dstl8 reads logs like a senior SRE, using sentiment & severity, generates clear cause-and-effect explanations and impact, backed by evidence. It’s like an extra observability engineer for your team.
- In-Context Troubleshooting: Dig deeper in context to issues and incidents. Pinpoint failures faster and reduce the effort to fix them. Surface patterns, trends, hot spots through dynamic heatmaps, timelines, and log details.
- Monitor. Distill. Explain: Powered by M�bius, a continuous, multi-layer AI that distills, correlates, investigates in real-time. Distills all your logs at the edge, detects anomalies in real time, with always-on agents that investigate, summarize, act, and learn from every incident.
VMblog: What are the biggest obstacles to Kubernetes adoption and scaling in 2025? How does your solution help organizations overcome these barriers?
Quillin: Kubernetes adoption isn’t slowing down, but in 2025, most teams are hitting the same wall: operational complexity, signal overload, escalating toil, and staff cutbacks. Clusters may be easy to spin up but hard to understand when things go haywire. The volume of telemetry produced including logs, traces, metrics, has outpaced the human ability to make sense of it. Add AI workloads and the debugging surface becomes unmanageable.
That’s where Dstl8, powered by M�bius continuous AI, steps in. Instead of sending every signal upstream, Dstl8 performs Edge Distillation analyzing telemetry locally to extract only the essential information. Its Correlation Layer connects patterns across clusters, and its Incident Layer explains what changed, why, and what to do next. By applying AI at the source and up to the incident layer, we help teams scale Kubernetes without scaling confusion or cost. The result: faster insight, quicker turns, and a clearer understanding of where the problem is and how to fix it, so engineers can focus on reliability not wrangling data.
VMblog: With platform engineering gaining momentum, how do you support organizations building internal developer platforms and improving developer experience?
Quillin: Platform engineering succeeds when developers get speed, consistency, simplicity and clarity, but too often, observability tools slow things down with confusing query languages, dashboards, and too much haystack, too little needle. Dstl8, powered by M�bius continuous AI, brings observability distillation directly into internal developer platforms. At the Edge Distillation layer, it filters and summarizes telemetry before it overloads developers and shared systems, giving. The Correlation Layer builds shared context across services so platform teams can expose actionable insights through self-service interfaces instead of raw data dumps. And the Incident Layer lays out clear guidance, dynamic developer-friendly visualizations, and sufficient detail to support and explain.
For developers, that means fewer dashboards, faster feedback, and clearer answers inside the tools they already use. For platform teams, it means less operational drag and a simpler, more explainable troubleshooting and find/fix experience.
VMblog: How is your company addressing the intensifying focus on cloud cost optimization and FinOps in cloud-native environments?
Quillin: Telemetry growth continues to spiral out of control. Dstl8 tackles it at the source by distilling all signals down to their essence before export, before ingest, and before storage. M�bius agents learn and adapt what data drives insight, cutting ingestion volume and cost while improving mean-time-to-resolution. FinOps friendly for sure.
VMblog: What’s your take on the convergence of security-by-design and cloud-native development in 2025? How do you help customers implement secure-by-default practices?
Quillin: Security and observability are converging fast as both depend on logs and understanding system behavior in real time, not after the fact. In 2025, “secure-by-design” means building platforms that see when something deviates from normal before it becomes an incident. That’s where Dstl8, powered by M�bius continuous AI, helps. By analyzing telemetry at the Edge Distillation layer, Dstl8 identifies unexpected changes, permission shifts, and access anomalies close to their source before they propagate. The Correlation Layer connects these signals across clusters, revealing patterns that might indicate drift or compromised dependencies. And at the Incident Layer, Dstl8 generates clear, explainable narratives that help engineers validate and remediate issues quickly. We’re not a security product per se, but Dstl8 builds secure-by-default observability, where clarity, traceability, and early anomaly detection become the first line of defense.
VMblog: What creative booth experiences, prize draws, or community engagement activities have you planned to connect with attendees?
Quillin: The ControlTheory booth will be all about fun and hands-on practical tools this year. Our booth will feature live Dstl8 demos showing how M�bius continuous AI distills real Kubernetes telemetry in real time, plus plenty of ways to get involved with our open-source project Gonzo.
Attendees can grab limited-edition Gonzo stress balls, Gonzo T-shirts, and other giveaways while trying out the latest features in both the open-source and enterprise platforms. We may have some major swag for a select few so come meet the team and see what AI-powered observability looks like in action. Whether you’re an SRE, developer, or platform engineer, you’ll leave our booth with something useful and probably a cool Gonzo giveaway in your hand.
VMblog: Looking ahead to 2026 and beyond, how do you see the cloud-native landscape transforming? What should enterprises be strategically planning for?
Quillin: We’re moving from observability to understanding – driven by distillation. AI agents are becoming first-class citizens in the SRE stack, not replacing humans, but amplifying them. Kubernetes, OpenTelemetry, and AI observability will converge into unified control planes that adapt in real time. ControlTheory’s role is to help teams navigate that shift, distilling complexity into clear, actionable intelligence.
VMblog: What emerging technologies or industry shifts is your company monitoring most closely as we head into 2026?
Quillin: We’re watching three major shifts that are reshaping cloud-native operations heading into 2026:
- Continuous AI: The next generation of infrastructure won’t rely on static models or reactive automation. It will be powered by AI that runs continuously, learning, correlating, and adapting in real time. That’s the foundation of our M�bius architecture. Just as CI/CD changed how we ship code, Continuous AI will change how we understand and operate systems.
- Edge Intelligence: As clusters scale and latency becomes critical, analysis is moving closer to the source. Running AI at the edge inside Kubernetes, not above it, will be essential for speed, cost, and control. Our Edge Distillation layer is designed for that shift.
- Converged Observability + FinOps + SecOps: The lines between operational, financial, and security visibility are blurring. The future will demand unified context across all signals, not separate stacks. M�bius continuous AI is already modeling those relationships dynamically across layers.
In short, we see a move from dashboards to continuous understanding: observability that’s alive, adaptive, and explainable. That’s where ControlTheory is focused heading into 2026: helping teams harness Continuous AI to turn complexity into clarity.
VMblog: What’s your top recommendation for attendees to maximize their KubeCon experience and learning outcomes?
Quillin: KubeCon can be overwhelming, in both great ways and tough ways. The trick is not to try to see everything but to connect the patterns, be intentional, and pace yourself. Look for sessions that bridge disciplines and focus on real customer use cases that include observability, AI, and platform engineering, because that’s where the next breakthroughs are happening. Our advice: talk to the founders (we’re there, say hi) and take time to ask how, what, and why. How can this help me? What can I learn and apply in my job and life when I head back home? There’s so much new to learn in the AI era, take the opportunity to educate yourself one sip at a time. Those conversations will point to the future.
And, of course, stop by our booth. You’ll see Continuous AI in action with Dstl8 and Gonzo and maybe walk away with a new way to think about clarity, control, monitoring, and observability (plus a Gonzo T-shirt)!
##
