Metoro Kubernetes Profiling
metoro.io
- Free plan
- Linux
- Web
- API
- Self-hosted
What Metoro Kubernetes Profiling is
Metoro Kubernetes Profiling continuously samples Kubernetes workloads with an eBPF profiler that runs at the kernel level. It creates interactive flame graphs that can be grouped by service, pod, or container, then filtered by cluster, namespace, deployment, or Kubernetes label. Metoro lists CPU and memory profiling, sampling at 97Hz per CPU core and a design target of less than 0.5% CPU overhead per node. Supported runtimes include Go, Python, C++, C, Rust, Node.js, and Java; other runtimes can appear through host-level frames. Profiling can include third-party containers and vendor workloads, without SDKs, code changes, agents, restarts, or rebuilds. 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 covers one cluster, one user, and two nodes with 28-day retention. Scale costs $20 per node per month, plus $0.20 per GB above 100GB per node. Enterprise pricing is by contract.
Who it suits
It suits teams investigating CPU or memory behavior across Kubernetes workloads, including third-party containers they do not build. It supports Linux-based Kubernetes environments; eBPF requires Linux kernel 4.18 or newer.
Good to know, and what to watch
- Interactive flame graphs can be grouped and filtered.
- Profiles connect with traces, logs, metrics, and deployment context.
- No SDKs, agents, restarts, or rebuilds are required.
- Free Hobby plan covers up to two nodes.
- eBPF requires Linux kernel 4.18 or newer.
- Hobby is limited to one cluster and one user.
- Scale adds $0.20 per GB above 100GB per node.
Plans
Every plan and priceBYOC
Enterprise
On Prem
Hobby
Free plan · 14-day trial
Getting started
- Open the Metoro website
- Choose Metoro Cloud, BYOC, or on-premises deployment
- Install Metoro with Helm
- Select the Hobby, Scale, or Enterprise plan
- Connect supported Kubernetes clusters and integrations
Limits to check
Hobby includes one cluster, one user, two nodes, and 28-day retention. Scale has 28-day retention and charges $0.20 per GB for usage above 100GB per node; Enterprise offers custom retention.
The Club's review
Metoro Kubernetes Profiling continuously profiles Kubernetes workloads with a kernel-level eBPF sampler. It is a strong fit for teams that need CPU and memory visibility across clusters, including vendor containers they cannot rebuild, without adding instrumentation to application code. The principal compromise is that serious multi-node deployments move quickly from the tightly capped free tier to per-node and excess-data charges.
Overview
Metoro collects profiles without SDKs, language agents, code changes, restarts, or rebuilds. That makes it practical for third-party workloads as well as services a team owns, where changing build pipelines is difficult or impossible. Sampling runs at 97Hz per CPU core and is designed for less than 0.5% CPU overhead per node, a sensible target for continuous observation.
CPU and memory profiles connect to Metoro APM traces, logs, metrics, pod state, and deployment context. This gives an operator a route from a hotspot to the affected workload and its operating conditions rather than an isolated flame graph. Hobby and Scale retain data for 28 days.
Key features
Kubernetes-focused flame graphs
Interactive flame graphs can be aggregated by service, pod, or container, then filtered by cluster, namespace, deployment, or Kubernetes label. Those dimensions are useful when a fleet-wide problem must be reduced to one workload, release, or environment.
Runtime and workload coverage
Supported runtimes include Go, Python, C++, C, Rust, Node.js, and Java. Host-level frames cover other runtimes, so the tool is not limited to applications written in the named languages. Because profiling happens with eBPF, teams can inspect vendor and other third-party containers they do not build themselves.
Operational workflow
Continuous alerts are supported, and GitHub integration uses a GitHub App that can open pull requests with suggested fixes and create GitHub issues from investigations. Metoro Cloud connects with Slack, PagerDuty, Datadog, Grafana, webhooks, REST APIs, and OpenTelemetry. These integrations make it more useful in an established on-call workflow than a profiler that stops at visualization.
Deployment and security
Installation uses Helm. Customers can choose Metoro Cloud, bring-your-own-cloud (BYOC), or on-premises deployment. Metoro Cloud states SOC 2 Type II certification, TLS 1.3 in transit, AES-256 at rest, SSO/SAML 2.0, role-based access control, and immutable audit logs. Scale support includes Slack access to Metoro engineers; Enterprise adds 24/7 white-glove support and onboarding.
Pricing
Hobby
Hobby is free forever at no cost. It covers one cluster, one user, and two nodes, with 28-day retention. This is suitable for an individual operator or a small evaluation, but the node and user ceilings make it a poor fit for a shared production observability program.
Scale
Scale costs $20 a month, billed at $20 USD / node / month, with $0.20 / GB on excess over 100GB / node. It allows unlimited clusters, users, and nodes while keeping 28-day retention. The expanded quotas fit growing teams, but per-node billing and overage charges should be modeled against both cluster size and profile volume before adoption. A 14-day trial provides a short window to validate that estimate.
Enterprise
Enterprise uses custom pricing under a contract and supports BYOC or on-premises deployment. It includes unlimited clusters, users, and nodes, with custom retention and custom SLAs. This is the appropriate tier for organizations with contractual requirements or retention needs beyond Scale; its price must be negotiated.
Platforms
Metoro offers API, Linux, self-hosted, and web access. Kubernetes support covers EKS, GKE, AKS, OpenShift, Rancher, k3s, kind, and self-managed CNCF-conformant distributions. The eBPF profiler requires Linux kernel 4.18 or newer, which is an important infrastructure prerequisite.
Who it's for
Choose Metoro when Kubernetes is central to operations and you need always-on CPU and memory profiling tied to traces, logs, metrics, and deployment state. It is particularly well suited to platform and SRE teams investigating third-party workloads or avoiding application rebuilds. Hobby suits a small proof of concept; Scale suits teams that need shared access across multiple clusters; Enterprise is for organizations requiring private deployment, custom retention, or contractual support.
Pros and cons
Pros
- No SDKs, agents, code changes, restarts, or rebuilds, including for containers you do not build.
- Kernel sampling is designed for less than 0.5% CPU overhead per node, supporting continuous collection.
- Flame-graph filters and links to operational telemetry help narrow production incidents quickly.
- Cloud, BYOC, and on-premises choices, plus stated enterprise security controls, accommodate different governance models.
Cons
- Hobby is limited to one cluster, one user, and two nodes, so it is mainly an evaluation tier.
- Scale's $20 USD / node / month pricing and $0.20 / GB excess charge can become material in large environments.
- Linux kernel 4.18 or newer is required for eBPF profiling.
Alternatives
For a broader catalog of profiling products, see Continuous Profiling Tools and eBPF Observability Tools.
- Guance Container Monitoring is a freemium option with Android, iOS, macOS, Windows, Linux, API, self-hosted, and web access, plus a free tier for up to 10 hosts.
- Middleware is a freemium observability platform whose free tier includes up to 100GB of data, 1,000 RUM sessions, and 20,000 synthetic checks.
- Coroot offers an Apache 2.0 self-hosted Community Edition and a Standard plan priced per monitored CPU.
- OneUptime is an Apache 2.0 open-source core platform that can be inspected, modified, and self-hosted.
- Zymtrace has a free Basic plan with CPU profiling for up to five concurrent machines and a GPU profiling trial.
- Amazon CodeGuru Profiler offers a 90-day free tier before standard sampling-hour pricing applies.
- Elastic Observability uses usage-based pricing, with logs and traces ingest starting at $0.09.
- Grafana Cloud Application Observability has an always-free tier with 2,232 host hours and usage-based Pro billing.
Verdict
Metoro is a strong choice for Kubernetes platform and SRE teams that want low-overhead, continuous CPU and memory profiling with deployment-aware context and no application instrumentation. Look elsewhere if the two-node Hobby ceiling is insufficient and per-node plus excess-data billing does not fit the environment, or if Linux kernel requirements cannot be met.
The verdict
Metoro combines low-overhead Kubernetes profiling with operational context and flexible deployment choices. Check the Hobby limits and Scale usage charges against the size of the environment.
Questions people ask
Is there a free plan?
Yes. Hobby is free forever and includes one cluster, one user, two nodes, and 28-day retention.
How much does Scale cost?
Scale is $20 per node per month, with $0.20 per GB for usage above 100GB per node.
Which deployment models are available?
Metoro is available as Metoro Cloud, BYOC, or an on-premises deployment.
Which Kubernetes environments are supported?
Supported environments include EKS, GKE, AKS, OpenShift, Rancher, k3s, kind, and self-managed CNCF-conformant distributions.
Does profiling require application changes?
No. It requires no SDKs, application code changes, language agents, restarts, or rebuilds.
What the maker publishes
- AI operations
- Metoro offers automated issue detection, alert investigation, root-cause analysis, deployment verification, runtime-to-code context and remediation workflows. source
- AI privacy
- The AI policy says AI is opt-in, providers do not train on customer data, connections and stored results are encrypted, and agents have no general internet access. source
- API
- The API reference documents a POST environment-creation endpoint using bearer authentication that returns an environment token. source
- APM
- Its eBPF collection maps Kubernetes services, requests, latency, errors and traces without manual instrumentation. source
- Automatic instrumentation
- Metoro installs an eBPF DaemonSet with Helm and gathers telemetry without application code changes, SDKs, sidecars or container restarts. source
- Company founding
- LinkedIn gives Metoro a 2023 founding year and San Francisco, California headquarters. source
- Company size
- LinkedIn lists Metoro as privately held with 11–50 employees. source
- Continuous profiling
- The Kubernetes Profiling feature continuously samples CPU activity at the kernel, attributes stacks to pods and services, and relates changes to deployments, traffic shifts and noisy neighbors. source
- Correlation
- Profiles are connected to Metoro APM traces, logs, metrics, pod state, and deployment context. source
- Deployment
- Metoro is installed with Helm and is available as Metoro Cloud, BYOC, or on-premises deployment. source
- Deployment options
- Metoro offers managed Cloud, BYOC in a customer account and on-premises deployment models. source
- Documentation and support
- Metoro provides Mintlify-hosted technical documentation, a community Slack and dedicated Slack Connect support for customers. source
- Flame graphs
- Interactive flame graphs can be aggregated by service, pod, or container and filtered by cluster, namespace, deployment, or Kubernetes label. source
- Flame graphs
- Interactive flame graphs can be aggregated by service, pod or container and filtered by cluster, namespace, deployment or Kubernetes label. source
- Funding
- Crunchbase lists a Metoro pre-seed funding round, but the amount and date are hidden on the public page. source
- Instrumentation
- Profiling requires no SDKs, application code changes, language agents, restarts, or rebuilds. source
- Integrations
- Metoro Cloud integrates with Slack, PagerDuty, Datadog, Grafana, webhooks, REST APIs, and OpenTelemetry. source
- Integrations
- Official documentation lists Slack, PagerDuty and webhooks as alert destinations, while the Cloud page also names Datadog and Grafana. source
- Kubernetes support
- Metoro states support for EKS, GKE, AKS, OpenShift, Rancher, k3s, kind, and self-managed CNCF-conformant Kubernetes distributions; eBPF requires Linux kernel 4.18 or newer. source
- Languages
- Metoro lists Go, Python, C++, C, Rust, Node.js, and Java as supported runtimes, with host-level frames for other runtimes. source
- Limit: G2 reviewer
- That reviewer said the main difficulty was making sense of the large amount of information presented. source
- Limit: independent review
- FunBlocks said regulated teams may be uneasy about AI-generated code changes and requested finer controls over agent decisions plus broader custom-CRD support. source
- Logs and events
- The platform searches application logs, Kubernetes events, deployment changes and workload metadata through one investigation path. source
- MCP and SDK extensibility
- Metoro publishes an MCP server that can be built in Go and exposes tools for querying Metoro data from AI clients, although it cannot start an AI investigation run directly. source
Checked against the maker's pages on 15 Oct 2026.