Caretta vs Metoro Kubernetes Profiling

Caretta is a free, self-hosted Kubernetes tool from groundcover that creates a visual map of services and their network interactions. It uses eBPF for collection and is designed for Linux systems with kernel 4.16 or later and CO-RE support. The project describes a minimal footprint and says no cluster modifications are required. Installation is documented through a Helm chart, with a recommendation to use a new, unique namespace. The chart includes Grafana with a predefined Caretta dashboard and VictoriaMetrics for collecting and publishing metrics that Prometheus-compatible dashboards can consume. Both bundled services are enabled by default, as is a five-second metrics polling interval, while Helm values can disable either component. Persistent storage for VictoriaMetrics starts disabled but can be enabled. Grafana’s default setup enables anonymous mode and includes administrator credentials in the README. Caretta is self-hosted, Apache-2.0 licensed, and support is directed through Slack or repository issues.

Metoro Kubernetes Profiling continuously profiles Kubernetes workloads with an eBPF sampler running at the kernel level. It creates interactive flame graphs that can be grouped by service, pod, or container and filtered by cluster, namespace, deployment, or Kubernetes label. Metoro lists Go, Python, C++, C, Rust, Node.js, and Java support, with host-level frames for other runtimes. Profiling does not require SDKs, code changes, language agents, restarts, or rebuilds, and can cover third-party containers and vendor workloads. Profiles connect to Metoro APM traces, logs, metrics, pod state, and deployment context. Installation uses Helm, with Cloud, BYOC, and on-premises deployment options. The free Hobby plan includes one cluster, one user, and two nodes with 28-day retention. Scale costs $20 USD per node per month, plus $0.20 per GB above 100GB per node. Enterprise pricing is contract-based and offers custom retention and SLAs. eBPF requires Linux kernel 4.18 or newer.