Skip to main content
Hong Kong
AIMenta
P

Pixie

by New Relic / CNCF

Auto-instrumentation observability platform for Kubernetes using eBPF to collect application traces, metrics, and logs without code changes or redeployment — instant APAC Kubernetes visibility.

Visit Pixie (opens in new tab) Get a recommendation
AIMenta verdict
Decent fit
4/5

"Auto-instrumented K8s observability — APAC platform teams use Pixie to instantly collect traces, metrics, and logs from APAC Kubernetes workloads using eBPF without adding instrumentation code or redeploying APAC services."

Features
6
Use cases
1
Watch outs
3
What it does

Key features

  • Zero-code instrumentation: eBPF auto-captures APAC HTTP, gRPC, SQL, DNS without SDK
  • Instant deployment: single DaemonSet install; APAC cluster visible in minutes
  • Application-layer traces: HTTP/gRPC request/response bodies via eBPF for APAC debugging
  • SQL query inspection: per-query APAC PostgreSQL/MySQL execution time via eBPF probes
  • PxL scripting: custom APAC telemetry queries against in-cluster data
  • Data-on-edge: raw APAC traces stay on nodes — only aggregates leave the cluster
When to reach for it

Best for

  • APAC platform teams who need instant Kubernetes observability without OpenTelemetry instrumentation overhead — particularly for debugging APAC incidents in services that lack manual instrumentation.
Don't get burned

Limitations to know

  • ! Short data retention (minutes to hours in memory) — APAC long-term analysis requires exporting to Prometheus/Tempo
  • ! Linux kernel 4.14+ required — APAC teams on older node OS versions cannot use Pixie
  • ! Less mature than established APAC APM vendors — PxL scripting curve for custom analysis
Context

About Pixie

Pixie is a Kubernetes observability platform (CNCF sandbox project, maintained by New Relic) that uses eBPF to automatically instrument APAC Kubernetes workloads — collecting application-level traces, CPU/memory metrics, network flows, and application logs without modifying APAC application code or adding sidecar containers. APAC platform teams deploy Pixie once as a DaemonSet and get instant observability across the entire APAC cluster.

Pixie's auto-instrumentation captures application-layer data: HTTP/gRPC request/response headers and bodies, PostgreSQL and MySQL query text and execution time, DNS queries, and JVM metrics for Java APAC workloads — all through eBPF probes at the kernel level. APAC teams get query-level database observability and service mesh-equivalent HTTP tracing without modifying a single APAC service.

Pixie's PxL scripting language allows APAC platform engineers to write custom queries against the telemetry data Pixie collects — similar to PromQL for metrics but covering traces and logs. APAC engineers can write PxL scripts to detect slowest SQL queries per APAC service, find HTTP 5xx patterns by endpoint, or identify which APAC pods are generating the most inter-node network traffic.

Pixie's data-on-edge architecture keeps raw APAC telemetry on the Kubernetes nodes — only aggregated query results leave the APAC cluster, addressing data sovereignty concerns for APAC regulated workloads. Short-term data (minutes to hours) is retained in node memory for APAC incident investigation without egressing raw traces to external storage.

Beyond this tool

Where this category meets practice depth.

A tool only matters in context. Browse the service pillars that operationalise it, the industries where it ships, and the Asian markets where AIMenta runs adoption programs.

Or browse All tools · Encyclopedia · Case studies · Rankings