Groundcover Expands eBPF Sensor With SQL Tracing
The update links slow application requests to the exact SQL queries behind them, adding Kubernetes context without code changes.
Updated on Nov. 9, 2026 in Data Centers
Groundcover expanded its eBPF sensor with SQL query tracing and Kubernetes context, announced at KubeCon North America 2026. The sensor also added ClickHouse and SIP protocol support.
Why it matters
The sensor collects telemetry from every Kubernetes cluster it runs in without requiring application code changes, which lets teams trace performance problems down to the database query level. Its ability to run in managed and locked-down clusters widens where such monitoring can be deployed.
The sensor links each request span to the exact SQL query it executes and tags data with workload, namespace, cloud region, and zone at collection. Metrics for processes blocked on I/O and Azure OpenAI time to first token are now included, though no measured overhead figures were disclosed.
The players
groundcover
A cloud-native observability company that builds eBPF-based monitoring tools for Kubernetes environments.
The details
The sensor now links application request spans to the SQL queries they execute, adds Kafka to the workload APIs view, and detects Azure OpenAI workloads. StatsD ingestion is on by default, and OpenTelemetry-only deployments receive full APM support.
Timeline
November 9, 2026: groundcover announced the sensor expansion at KubeCon North America 2026.
The Tech Race
The expansion extends eBPF-based observability, a kernel-level instrumentation approach that has gained adoption as an alternative to per-application SDKs. Adding SQL query linkage moves that model further into application-layer diagnostics that traditionally required code-level agents.
Engineers running Kubernetes clusters can trace slow requests to specific database queries without modifying application code or adding database client libraries. Teams using managed or locked-down clusters, including Google Kubernetes Engine Autopilot, can deploy the sensor as an approved workload.
The takeaway
Teams evaluating observability tooling can now assess eBPF-based SQL tracing as a way to diagnose database-driven latency without instrumenting each service. Checking whether a given sensor supports your cluster's security posture, such as Autopilot or proxy-routed traffic, should come before adoption.
Further reading
For more on the infrastructure these monitoring tools run on, see Data Centers.
