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: cilium-ebpf-host-routing-mtu-path-discovery-guardrail

Cilium eBPF Host-Routing MTU & Path Discovery Production Guardrail

Enforces strict interface MTU calculation, MSS clamping, and explicit ICMP Type 3 Code 4 handling across Cilium eBPF host-routing configurations. Eliminates silent blackholing, packet drops, and path fragmentation induced by Geneve/VXLAN overlay encapsulation and Direct Server Return (DSR) headers in high-throughput node topologies. Prevents asymmetric bypass of the Linux TCP/IP stack where eBPF drops Path MTU Discovery (PMTUD) signaling.

Prohibited Anti-Pattern: Silent Overlay Packet Blackholing via Unadjusted Native MTU and Broken eBPF PMTUD

Export Ready-to-Drop Configuration

Select format below to copy code or download file directly.

---
description: Cilium eBPF Host-Routing MTU & Path Discovery Production Guardrail - Enforces strict interface MTU calculation, MSS clamping, and explicit ICMP Type 3 Code 4 handling across Cilium eBPF host-routing configurations. Eliminates silent blackholing, packet drops, and path fragmentation induced by Geneve/VXLAN overlay encapsulation and Direct Server Return (DSR) headers in high-throughput node topologies. Prevents asymmetric bypass of the Linux TCP/IP stack where eBPF drops Path MTU Discovery (PMTUD) signaling.
globs: helm/**/*.{yaml,yml}, kubernetes/**/*.{yaml,yml}, infra/**/*.{yaml,yml}, terraform/**/*.tf, cilium-values.yaml
alwaysApply: true
verified_by: "agentrules.org"
license: "CC-BY-4.0"
source: "https://agentrules.org/rules/cilium-ebpf-host-routing-mtu-path-discovery-guardrail"
---

# Cilium eBPF Host-Routing MTU & Path Discovery Production Guardrail

> **Severity**: CRITICAL | **Category**: devops
> **Anti-Pattern Warning**: Silent Overlay Packet Blackholing via Unadjusted Native MTU and Broken eBPF PMTUD

## System Constraints
You are a Principal Infrastructure & Network Reliability Engineer enforcing the Cilium eBPF Host-Routing MTU Path Discovery Guardrail. Mandate invariants that prevent packet blackholing, silent TCP segmentation faults, and PMTUD failure caused by encapsulation overhead (VXLAN/Geneve) during eBPF host-routing. You must strictly validate Helm values, DaemonSet configurations, and CNI manifest definitions to ensure that overlay MTU deductions, TCP MSS clamping, and route-table parity are deterministically declared.

## Strict Guidelines
- Engineers MUST ALWAYS explicitly compute and define Cilium's MTU based on the physical interface MTU minus the exact overlay encapsulation overhead (e.g., physical MTU 1500 minus 50 bytes for Geneve = 1450 MTU; minus 60 bytes for WireGuard/IPsec when encrypted).
- Engineers MUST NOT rely on Cilium automatic MTU discovery in heterogeneous network topographies or public cloud virtual fabrics (AWS/GCP/Azure) where underlying NIC MTUs can trigger inaccurate local route caching.
- Engineers MUST enable TCP MSS clamping (`bpf.masquerade=true` and `bpf.hostRouting=true` accompanied by `--enable-pmtu-discovery=true` and `--enable-tcx=true` where supported) to preemptively negotiate packet limits before eBPF packet redirect bypasses kernel fragmentation.
- Manifests MUST configure ICMP Fragmentation Needed (Type 3, Code 4) ingress allowances across all network security groups and eBPF filters, preventing blackhole scenarios when Don't Fragment (DF) flags are asserted.
- Engineers MUST ensure cilium host-routing mode (`bpf.hostRouting: true`) is matched with `autoDirectNodeRoutes: true` on identical L2/L3 broadcast domains to avoid unnecessary VXLAN/Geneve encapsulation overhead for intra-rack traffic.

## Bad vs Good Implementations
### Cilium Helm Configuration: Automated Flawed MTU vs Explicit Deterministic Sizing

#### Bad Antipattern (AI Common Hallucination)
> **Why it fails**: Failing to explicitly declare the tunnel MTU deductions (50B Geneve overhead on 1500 byte links) causes pods to inherit a 1500 byte MTU. High-throughput TCP streams send full 1500-byte payloads with DF set, which eBPF host routing attempts to encapsulate into 1550-byte frames. The physical NIC or intermediate switches drop the frame without returning an ICMP fragmentation message, producing persistent connection hangs (blackholing).
```yaml
# INSUFFICIENT: Implicit MTU and host routing without MSS clamping or PMTUD handling
cilium:
  routingMode: tunnel
  tunnelProtocol: geneve
  bpf:
    masquerade: true
    hostRouting: true
  # Missing explicit mtu field: defaults to physical 1500 or flawed auto-detection
  # Missing PMTUD enablement: packets exceeding 1450 will be silently dropped in eBPF
```

#### Production Standard (Correct)
> **Standard**: Explicitly declares an MTU of 1450, reserving 50 bytes for Geneve encapsulation overhead. Enables Path MTU Discovery and autoDirectNodeRoutes to avoid encapsulation on contiguous subnets, while guaranteeing that eBPF host-routing honors MSS clamping and drops packets within pod namespaces safely before reaching physical interfaces.
```yaml
# CORRECT: Fully reconciled MTU, explicit Geneve overhead, and enabled PMTUD
cilium:
  routingMode: tunnel
  tunnelProtocol: geneve
  autoDirectNodeRoutes: true
  mtu: 1450 # 1500 (Physical MTU) - 50 (Geneve Encap)
  bpf:
    masquerade: true
    hostRouting: true
  pmtuDiscovery:
    enabled: true
  ipv4:
    enabled: true
  # Ensure TCP MSS clamping is applied to synthesized pod connections
  bpfMasquerade:
    enabled: true
```
Recommended MCP Servers (mcp-serverlist.com)

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

Kubernetes Cluster Inspector MCP
Inspects live DaemonSet configurations, node interface MTUs, and Cilium agent ConfigMaps across active clusters.
Recommended Agent Skills (askills.dev)

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

Cilium CLI Node Connectivity Diagnoser
Runs automated MTU path validation and encapsulation packet loss tests using `cilium connectivity test`.
Architectural Specification & Invariant Rationale

Architectural Invariant Design: Cilium eBPF Host-Routing MTU & Path Discovery Production Guardrail

Cilium eBPF Host-Routing MTU & Path Discovery Production Guardrail defines an industrial-grade behavioral standard for devops engineering. Enforces strict interface MTU calculation, MSS clamping, and explicit ICMP Type 3 Code 4 handling across Cilium eBPF host-routing configurations. Eliminates silent blackholing, packet drops, and path fragmentation induced by Geneve/VXLAN overlay encapsulation and Direct Server Return (DSR) headers in high-throughput node topologies. Prevents asymmetric bypass of the Linux TCP/IP stack where eBPF drops Path MTU Discovery (PMTUD) signaling. 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: Silent Overlay Packet Blackholing via Unadjusted Native MTU and Broken eBPF PMTUD 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 DriftSilent Overlay Packet Blackholing via Unadjusted Native MTU and Broken eBPF PMTUDCritical
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 explicitly compute and define Cilium's MTU based on the physical interface MTU minus the exact overlay encapsulation overhead (e.g., physical MTU 1500 minus 50 bytes for Geneve = 1450 MTU

Engineers MUST ALWAYS explicitly compute and define Cilium's MTU based on the physical interface MTU minus the exact overlay encapsulation overhead (e.g., physical MTU 1500 minus 50 bytes for Geneve = 1450 MTU; minus 60 bytes for WireGuard/IPsec when encrypted).

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 rely on Cilium automatic MTU discovery in heterogeneous network topographies or public cloud virtual fabrics (AWS/GCP/Azure) where underlying NIC MTUs can trigger inaccurate local route caching.

Engineers MUST NOT rely on Cilium automatic MTU discovery in heterogeneous network topographies or public cloud virtual fabrics (AWS/GCP/Azure) where underlying NIC MTUs can trigger inaccurate local route caching.

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

Engineers MUST enable TCP MSS clamping (`bpf.masquerade=true` and `bpf.hostRouting=true` accompanied by `--enable-pmtu-discovery=true` and `--enable-tcx=true` where supported) to preemptively negotiate packet limits before eBPF packet redirect bypasses kernel fragmentation.

Engineers MUST enable TCP MSS clamping (`bpf.masquerade=true` and `bpf.hostRouting=true` accompanied by `--enable-pmtu-discovery=true` and `--enable-tcx=true` where supported) to preemptively negotiate packet limits before eBPF packet redirect bypasses kernel fragmentation.

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

Manifests MUST configure ICMP Fragmentation Needed (Type 3, Code 4) ingress allowances across all network security groups and eBPF filters, preventing blackhole scenarios when Don't Fragment (DF) flags are asserted.

Manifests MUST configure ICMP Fragmentation Needed (Type 3, Code 4) ingress allowances across all network security groups and eBPF filters, preventing blackhole scenarios when Don't Fragment (DF) flags are asserted.

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

Engineers MUST ensure cilium host-routing mode (`bpf.hostRouting

Engineers MUST ensure cilium host-routing mode (`bpf.hostRouting: true`) is matched with `autoDirectNodeRoutes: true` on identical L2/L3 broadcast domains to avoid unnecessary VXLAN/Geneve encapsulation overhead for intra-rack traffic.

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 Cilium eBPF Host-Routing MTU & Path Discovery Production Guardrail critical for AI coding assistants?

Enforces strict interface MTU calculation, MSS clamping, and explicit ICMP Type 3 Code 4 handling across Cilium eBPF host-routing configurations. Eliminates silent blackholing, packet drops, and path fragmentation induced by Geneve/VXLAN overlay encapsulation and Direct Server Return (DSR) headers in high-throughput node topologies. Prevents asymmetric bypass of the Linux TCP/IP stack where eBPF drops Path MTU Discovery (PMTUD) signaling. 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 (Silent Overlay Packet Blackholing via Unadjusted Native MTU and Broken eBPF PMTUD) 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

Cilium