Austin, TX, December 11, 2025 — ControlTheory is proud to announce early availability of Dstl8, a continuous AI platform that turns raw telemetry into real answers for developers and SRE teams. Built on the momentum of Gonzo, the company’s fast-growing open source log analysis tool, Dstl8 pushes observability toward a new model: distillation at the edge, correlation across environments, and always-on investigation by AI agents.
Developers and operators are facing the same problem across cloud-native systems: too many logs, too much noise, and too little time to make sense of any of it. Most teams still depend on collecting and storing everything, which slows investigations, drives up cost, and strains lean ops groups. ControlTheory’s approach flips that model by applying AI where telemetry is produced, not after the fact.
ControlTheory CEO, Bob Quillin notes that “Teams are drowning in telemetry. They’re being asked to ship more, observe more, and respond faster, but the old model of collecting everything is breaking. Gonzo showed the world that log analysis can be fast, simple, and in the flow of real work.” Now available in preview, “Dstl8 takes that momentum and turns it into continuous intelligence across environments. We aim to give every engineer a system that explains what changed and why, without extra toil or headcount.”
The Rise of Gonzo: From TUI Curiosity to Daily Driver
Gonzo launched in August 2025 as a terminal-native, OTLP-aware log analysis tool. As an easy to use TUI (terminal user interface). It quickly became part of the daily workflow for thousands of engineers who wanted fast, clear answers without dashboards or complex query languages.
Engineers embraced Gonzo because it matched how they already work:
- Group logs by pattern
- Scan for signals, not volume
- Track severity and hot spots in real time
- Use AI to suggest possible causes, locally or with preferred models
The community response confirmed this direction. Gonzo passed 2,000 GitHub stars in two months and continues to show strong adoption among Kubernetes and OpenTelemetry users.
“We built Dstl8 after watching how teams actually debug. Developers group logs, hunt for patterns, and chase down the signals that matter. Gonzo made that workflow instant in the terminal.” says Jonathan Reeve, ControlTheory Founder and CPO, “Dstl8 applies the same ideas everywhere across clusters, pipelines, and regions. The platform distills your logs at the edge, correlates what matters, and uses agents to explain what broke in plain language. It’s the system we always wished existed when we were on-call.”
Introducing Dstl8: Continuous AI Across All Your Logs
Dstl8 is the next step, built for teams who need system-wide clarity, not just single-pod views. It uses ControlTheory’s Möbius continuous AI to analyze logs at three layers:
Edge Distillation
Telemetry is summarized where it’s produced. Dstl8 cuts noisy log traffic by 90+% while keeping severity, sentiment, and key language-based signals intact.
This matters because logs are written in human language, not metrics. Dstl8 reads that language directly. It captures the tone, intent, and direction of change across services so engineering teams can spot rising frustration, cascading failures, or customer-impacting errors before they escalate.
Operational Inference Layer
Dstl8 connects patterns across clusters, services, and environments. It finds related log flows without requiring dashboards or saved queries.
Continuous Agentic Layer
Always-on agents investigate changes, explain what broke and why, and provide next-step guidance in plain language. These insights can be delivered through Slack, webhook, or the Dstl8 UI.
Dstl8 works alongside existing systems like CloudWatch, Datadog, Loki, and Elastic. It distills the noise before those tools ingest it, reducing cost while increasing signal quality.
Dstl8 also exposes insights through MCP, enabling developers to access correlations, summaries, and incident explanations directly from their IDE, Claude, Cursor, or any MCP-aware tool. Developers don’t have to switch context; Dstl8 meets them where they already work.
Why It Matters Now
Teams running Kubernetes, OpenTelemetry, and AI workloads are hitting the same wall: telemetry volume is growing faster than SRE bandwidth. Ops teams want clarity without scaling headcount. Developers want answers without learning another query language or maintaining more dashboards.
Dstl8 addresses these realities by giving teams a control plane built for continuous understanding, not just data collection.
Early users report:
- 90%+ lower telemetry volume transported and stored, without losing context
- Faster triage through correlation summaries generated in seconds
- Clearer communication across engineering and management
Dstl8 builds on the same workflow that made Gonzo popular: keep the developer in flow, reduce noise, and surface the right signals automatically.
The Observability Progression: Gonzo –> Dstl8
ControlTheory sees Gonzo and Dstl8 as part of one journey.
- Gonzo is immediate, hands-on visibility for individual developers.
- Dstl8 is continuous AI for teams, spanning dev, staging, and production.
Both share the same principles:
- distill first, investigate continuously
- treat signal quality as a first-class concern
- meet engineers where they already work (terminal, IDE, Slack)
Availability
Dstl8 entered full public preview at KubeCon + CloudNativeCon North America 2025. Thank you to all of the design partners who have collaborated on the initial release. Also, a special thank you to everyone at KubeCon who were instrumental in validating the value that Gonzo and Dstl8 bring to modern observability.
Get Started Today!
- Run Gonzo today: https://github.com/control-theory/gonzo
- Sign up for Dstl8: https://www.controltheory.com/product/dstl8/
Media Contact:
Organization: ControlTheory
Website: https://www.controltheory.com/
