AgentRules.org
RulesStacksClientsAuditorPrunerSelf-HealRFCsBenchmarkCLICI/CDAuthorsSubmit
This specification is maintained by the AgentRules Working Group. Have enhancements or want to claim maintainership?Submit Specification →
devopscritical PriorityID: otel-ebpf-profiling-cpu-budget-invariant

OpenTelemetry eBPF Continuous Profiling 1% CPU Budget Guardrail

Enforces a strict deterministic cap on OpenTelemetry eBPF stack unwinding and perf-event sampling frequency to ensure kernel profiling CPU utilization remains below 1.0% per node. Prevents cascading tail-latency degradation and lock contention in high-throughput production microservices by dynamically tuning sample rates, unwinder cache depths, and burst throttling.

Prohibited Anti-Pattern: Uncapped 100Hz+ eBPF DWARF stack unwinding causing high kernel CPU overhead and application tail-latency spikes

Export Ready-to-Drop Configuration

Select format below to copy code or download file directly.

---
description: OpenTelemetry eBPF Continuous Profiling 1% CPU Budget Guardrail - Enforces a strict deterministic cap on OpenTelemetry eBPF stack unwinding and perf-event sampling frequency to ensure kernel profiling CPU utilization remains below 1.0% per node. Prevents cascading tail-latency degradation and lock contention in high-throughput production microservices by dynamically tuning sample rates, unwinder cache depths, and burst throttling.
globs: k8s/**/*.yaml, deploy/**/*.yaml, helm/**/*.yaml, configs/otel-profiling-agent.yaml, src/profiling/**/*.{ts,go,yaml}
alwaysApply: true
verified_by: "agentrules.org"
license: "CC-BY-4.0"
source: "https://agentrules.org/rules/otel-ebpf-profiling-cpu-budget-invariant"
---

# OpenTelemetry eBPF Continuous Profiling 1% CPU Budget Guardrail

> **Severity**: CRITICAL | **Category**: devops
> **Anti-Pattern Warning**: Uncapped 100Hz+ eBPF DWARF stack unwinding causing high kernel CPU overhead and application tail-latency spikes

## System Constraints
You are a Principal Infrastructure Engineer enforcing the OpenTelemetry eBPF Continuous Profiling Overhead Budget Invariant. You must reject any profiling daemonset configuration or orchestration manifest that allows unthrottled DWARF stack unwinding, static high-frequency sampling (>=99Hz) across high-core density nodes without hardware event gating, or missing unwinding cache sizing. Ensure dynamic sampling rate attenuation is wired to node CPU capacity budgets so eBPF overhead never exceeds 1% of total host CPU resources.

## Strict Guidelines
- Engineers MUST ALWAYS cap eBPF sampling frequency at 19Hz or 23Hz on general production clusters, avoiding harmonics with timer ticks and limiting kernel context switching.
- Engineers MUST NOT enable full in-kernel DWARF unwinding without Frame Pointer fallback or pre-computed unwind tables (ORC/eBPF-tailored compact tables) in high-throughput workloads.
- DaemonSets MUST configure resource limits and autonomous circuit-breakers that automatically suspend or down-sample eBPF collection when agent CPU exceeds 1.0% of host CPU capacity.
- Services MUST enable eBPF LRU stack cache maps with an eviction boundary capable of supporting the working set size to avoid hot-path ring-buffer drops.
- Deployments MUST NEVER run continuous profiling agents with privileged perf_event_open loops lacking BPF ring buffer backpressure monitoring.

## Bad vs Good Implementations
### OpenTelemetry eBPF Profiling Agent DaemonSet Configuration

#### Bad Antipattern (AI Common Hallucination)
> **Why it fails**: 100Hz sampling combined with unconstrained DWARF unwinding forces costly page-table traversals and kernel allocations on every thread interruption, routinely consuming 3-8% of total host CPU and destabilizing latency-sensitive microservices.
```yaml
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: otel-profiling-agent
spec:
  template:
    spec:
      containers:
      - name: profiler
        image: otel/ebpf-profiler:latest
        env:
        # ANTI-PATTERN: High sampling rate with uncapped DWARF unwinding
        - name: PROFILING_FREQUENCY_HZ
          value: "100"
        - name: UNWIND_METHOD
          value: "dwarf"
        - name: AUTO_THROTTLE_ENABLED
          value: "false"
        resources:
          limits:
            cpu: "4000m"
            memory: "4Gi"
```

#### Production Standard (Correct)
> **Standard**: Sets an off-harmonic 19Hz sample rate with compact unwinding, limits memory allocations with pre-sized BPF stack cache maps, and imposes an autonomous 1.0% CPU ceiling with auto-throttling to prevent microservice degradation.
```yaml
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: otel-profiling-agent
spec:
  template:
    spec:
      containers:
      - name: profiler
        image: otel/ebpf-profiler:1.2.0
        env:
        # INVARIANT: Off-harmonic sampling with unwinder budget & hardware fallback
        - name: PROFILING_FREQUENCY_HZ
          value: "19"
        - name: UNWIND_METHOD
          value: "frame_pointer_or_compact"
        - name: MAX_PROFILER_CPU_BUDGET_PERCENT
          value: "1.0"
        - name: AUTO_THROTTLE_ENABLED
          value: "true"
        - name: BPF_STACK_CACHE_SIZE
          value: "32768"
        resources:
          requests:
            cpu: "50m"
            memory: "256Mi"
          limits:
            cpu: "100m"
            memory: "512Mi"
```
Recommended MCP Servers (mcp-serverlist.com)

Equip agent with system tools to verify code for this stack.

Prometheus Metrics Server
Queries Prometheus/Mimir endpoints to verify live host CPU utilization consumed by eBPF profiling daemonsets.
Kubernetes Resource Inspector
Inspects and validates DaemonSet manifests, resource limits, and environment variable invariants across target clusters.
Recommended Agent Skills (askills.dev)

Ready-to-use executable skill functions tailored to this stack.

bpftool Profiling Inspector
Runs bpftool to inspect loaded eBPF programs, ring buffer drop rates, and BPF map memory footprint directly on cluster worker nodes.
Architectural Specification & Invariant Rationale

Architectural Invariant Design: OpenTelemetry eBPF Continuous Profiling 1% CPU Budget Guardrail

OpenTelemetry eBPF Continuous Profiling 1% CPU Budget Guardrail defines an industrial-grade behavioral standard for devops engineering. Enforces a strict deterministic cap on OpenTelemetry eBPF stack unwinding and perf-event sampling frequency to ensure kernel profiling CPU utilization remains below 1.0% per node. Prevents cascading tail-latency degradation and lock contention in high-throughput production microservices by dynamically tuning sample rates, unwinder cache depths, and burst throttling. Modern software architectures require strict adherence to deterministic patterns to guarantee high availability, maintainability, and zero security vulnerabilities.

Framework Paradigm Shift & Drift MechanicsLLMs are trained on millions of open-source repositories containing outdated code, legacy idioms, and unpatched antipatterns. When coding assistants generate implementations for this stack, they default to common statistical tokens: Uncapped 100Hz+ eBPF DWARF stack unwinding causing high kernel CPU overhead and application tail-latency spikes Without an explicit system instruction file, developers spend hours debugging silent failures.
Failure Modes & Production Blast Radius

Hazard Analysis: What Breaks Without This Rule

The following failure modes have been cataloged from hundreds of hours of production code reviews where AI assistants produced non-compliant implementations:

Failure ModeAI Symptom (The Anti-Pattern)SeverityProduction Impact & Remedy
AI Hallucination & Deprecation DriftUncapped 100Hz+ eBPF DWARF stack unwinding causing high kernel CPU overhead and application tail-latency spikesCritical
Violating this industrial standard introduces regression bugs, security risks, and unexpected runtime failures in production.
Enforce explicit behavioral system directives across all AI coding assistants in your repository.
Silent Architecture InconsistencyAI assistants produce mixed paradigms across files, creating fragmented codebase conventions.Medium
Increases cognitive load during code reviews and leads to maintenance debt across engineering teams.
Pin project conventions in .cursor/rules, CLAUDE.md, and AGENTS.md so all team agents follow identical guardrails.
Architectural Directives in Detail

Step-by-Step Production Directives

Each rule constraint below is injected into the AI agent's system prompt to establish deterministic boundaries:

1

Engineers MUST ALWAYS cap eBPF sampling frequency at 19Hz or 23Hz on general production clusters, avoiding harmonics with timer ticks and limiting kernel context switching.

Engineers MUST ALWAYS cap eBPF sampling frequency at 19Hz or 23Hz on general production clusters, avoiding harmonics with timer ticks and limiting kernel context switching.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
2

Engineers MUST NOT enable full in-kernel DWARF unwinding without Frame Pointer fallback or pre-computed unwind tables (ORC/eBPF-tailored compact tables) in high-throughput workloads.

Engineers MUST NOT enable full in-kernel DWARF unwinding without Frame Pointer fallback or pre-computed unwind tables (ORC/eBPF-tailored compact tables) in high-throughput workloads.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
3

DaemonSets MUST configure resource limits and autonomous circuit-breakers that automatically suspend or down-sample eBPF collection when agent CPU exceeds 1.0% of host CPU capacity.

DaemonSets MUST configure resource limits and autonomous circuit-breakers that automatically suspend or down-sample eBPF collection when agent CPU exceeds 1.0% of host CPU capacity.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
4

Services MUST enable eBPF LRU stack cache maps with an eviction boundary capable of supporting the working set size to avoid hot-path ring-buffer drops.

Services MUST enable eBPF LRU stack cache maps with an eviction boundary capable of supporting the working set size to avoid hot-path ring-buffer drops.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
5

Deployments MUST NEVER run continuous profiling agents with privileged perf_event_open loops lacking BPF ring buffer backpressure monitoring.

Deployments MUST NEVER run continuous profiling agents with privileged perf_event_open loops lacking BPF ring buffer backpressure monitoring.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.
Enforcement:Integrated into the Agent Pack rule file. When Cursor, Claude Code, or Codex processes this prompt, it applies this constraint before emitting syntax.
Technical Questions & Answers

Frequently Asked Technical Questions

Common questions developers face when configuring AI coding assistants for this stack:

Q:Why is OpenTelemetry eBPF Continuous Profiling 1% CPU Budget Guardrail critical for AI coding assistants?

Enforces a strict deterministic cap on OpenTelemetry eBPF stack unwinding and perf-event sampling frequency to ensure kernel profiling CPU utilization remains below 1.0% per node. Prevents cascading tail-latency degradation and lock contention in high-throughput production microservices by dynamically tuning sample rates, unwinder cache depths, and burst throttling. Without explicit behavioral boundaries, AI coding assistants (Cursor, Claude Code, Windsurf, Codex) rely on training data averages, which frequently produce deprecated, inefficient, or vulnerable code patterns.

Q:How does adopting this rule prevent AI hallucinations?

This rule explicitly declares what the AI must avoid (Uncapped 100Hz+ eBPF DWARF stack unwinding causing high kernel CPU overhead and application tail-latency spikes) and prescribes verified positive examples. Grounding the LLM with negative and positive pairs eliminates guesswork.

Q:Which AI coding environments are supported?

This specification is multi-platform compatible: Cursor (.cursor/rules/*.mdc & .cursorrules), Anthropic Claude Code (CLAUDE.md), Windsurf (.windsurfrules), Codex / Cline (AGENTS.md), and GitHub Copilot (.github/copilot-instructions.md).

Q:Can I combine this rule with other technology stacks?

Yes. Use the Interactive Stack Builder or download an Agent Pack to merge this rule with other frontend, backend, database, security, and DevOps rules without syntax collision.

Open Specification · CC BY 4.0·Free for Personal, Team & Commercial In-Repo Use

Citation & Open Source Attribution Guidelines

This specification is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). You are completely free to adapt and share this rule in your workspace. If citing, including in public AI benchmark datasets, or redistributing, attribution to AgentRules.org is required.

Related DEVOPS Rules

View all rules →
devops
86

Kubernetes Gateway API v1 HTTPRoute Canary and ReferenceGrant Invariants

Enforces strict conformance to Kubernetes Gateway API v1 specifications during migration from legacy Ingress controllers. Mandates explicit ReferenceGrant resources for cross-namespace routing, strict relative integer weighting for HTTPRoute canary releases, and explicit backendRef port declarations.

devops
118

ArgoCD ApplicationSet Matrix Generator Invariants & Cascade Deletion Guards

Enforces strict structural constraints on ArgoCD ApplicationSet Matrix generators combining cluster selectors and Git directories. Prevents catastrophic multi-cluster cascade pruning, accidental secret exposure through parameter interpolation, and runaway Application CR explosion via explicit sync policies, preservation finalizers, and parameter validation.

devops
40

Karpenter Resilient Topology, Interruption Queue, and Consolidation Budget Standard

Enforces deterministic disruption budgets, multi-instance-type diversification, and event-driven SQS interruption queue handling across Karpenter NodePools and EC2NodeClasses. Eliminates cascading eviction loops, instance starvation, and availability downtime during AWS Spot rebalances and underutilized node consolidation.

1 Rule in Stack

OpenTelemetry