groundcover: AI-era observability with EBPF and agentic monitoring

Description

groundcover CEO Shahar Azulay discusses the company's full-stack observability platform built for the AI era. The platform uses EBPF technology to collect high-fidelity telemetry without instrumentation and operates on a bring-your-own-cloud architecture to keep data private and scalable. Azulay announces the GA of groundcover's AI agent mode, which enables autonomous troubleshooting and communication between coding agents and production observability data. The platform addresses unique challenges of monitoring AI workloads, including long-running agent sessions with tens of thousands of spans, token usage tracking, and hallucination detection. Key topics covered: • Full-stack observability platform architecture • EBPF technology for agentless telemetry collection • Bring-your-own-cloud AI operations model • AI agent mode for autonomous troubleshooting • Monitoring AI workloads and LLM integrations • Agent drift, evaluation, and governance • Migration strategies from legacy observability vendors • Future of observability in AI-driven development Chapters: 0:12 - Introduction to groundcover 0:34 - AI agent GA announcement 0:52 - EBPF and telemetry collection 1:05 - Bring your own cloud architecture 5:43 - AI observability challenges 6:23 - EBPF for AI governance 6:59 - Agent drift and evaluation 11:58 - Migration from legacy platforms Keywords: groundcover, observability, EBPF, AI agents, OpenTelemetry, cloud native, Kubernetes, telemetry, agentic AI, LLM monitoring, bring your own cloud