Observability

Sidecar-free Tracing

Using eBPF to answer the Service Mesh challenge

Tracing without sidecars

  • Traditional Network Observability

    • Tracing is essential for cloud native app operations
    • Current approaches involve running sidecar proxies in each workload or pod.
    • This has massive impact on latency, cost of operation and architectural complexity.

  • Cilium Network Observability

    • Native eBPF-based HTTP, DNS, & TLS visibility, enabling L7 tracing
    • High performance, in-Kernel eBPF protocol parsers
    • No need to inject proxies and introduce corresponding added latency
    • Visibility exported in Hubble, OpenTelemetry or in SIEM compatible formats
    • Preserve low latency, keep the cost of observability low

Tracing without the added overhead

  • 4x better latency than sidecar-based solutions.
  • More efficient resources utilization.
  • Less complex architecture, critically simplifying operations and maintenance.
  • Support for non-TCP protocols like UDP.

The Inside Track on Sidecarless Service Mesh

Get the latest on the sidecarless Cilium Service Mesh from Liz Rice and Thomas Graf!

Want to learn more?

  • Start a Lab

    Step through our interactive labs in a sandbox environment.

  • Contact sales

    Engage with the Isovalent sales team to assess an enterprise-grade deployment of Cilium through the Isovalent Enterprise Platform.

  • Get a demo

    Request a personalized live demo with an Isovalent Enterprise Platform expert.