storm

Diagnose reactive storms causing MeshWeaver portal outages from Kubernetes pod logs.

11|4|Updated Jan 15, 2024
One-click install
npx skills add https://github.com/Systemorph/MeshWeaver --skill storm-systemorph
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: storm
Source: https://github.com/Systemorph/MeshWeaver/tree/main/.claude/skills/storm
Command: npx skills add https://github.com/Systemorph/MeshWeaver --skill storm-systemorph

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill identifies and resolves reactive storms—unbounded loops or deadlocks that cause portal pods to hang, trigger liveness probe failures, and result in 502 outages.

Core Features & Use Cases

  • Root Cause Analysis: Distinguishes between sync-over-async deadlocks and uncaught exception loops.
  • Diagnostic Procedures: Provides specific CLI commands to analyze pod logs, identify silence windows, and track repeating tokens.
  • Use Case: When a portal pod repeatedly restarts with exit code 137, use this Skill to pinpoint the exact reactive stream or swallowed exception causing the accumulation.

Quick Start

Use the storm skill to diagnose the root cause of the recent 502 outage by analyzing the pod logs for repeating error patterns.

Frequently Asked Questions about storm

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I diagnose a reactive programming deadlock causing Kubernetes pod restarts?

Diagnose reactive programming deadlocks by analyzing Kubernetes pod logs for sync-over-async blocking patterns and silence windows. This pinpoints the exact reactive stream causing portal pods to hang and trigger liveness probe failures.

Why does my portal pod repeatedly restart with exit code 137 and cause 502 outages?

Portal pods restart with exit code 137 due to unbounded reactive loops or swallowed exception accumulation. Analyze repeating error tokens and reactive stream telemetry to identify the uncaught exception loop causing the outage.

What is the difference between a sync-over-async deadlock and an uncaught exception loop in reactive streams?

A sync-over-async deadlock blocks the actor-model message hub entirely, while an uncaught exception loop creates unbounded recursive failures. Both cause portal wedging, but require distinct diagnostic procedures to resolve.

How do I track swallowed exceptions in an actor-model message hub?

Track swallowed exceptions in an actor-model message hub by enforcing error propagation boundaries and analyzing reactive stream telemetry. Use specific CLI commands to identify repeating tokens and silence windows in pod logs.

Can I use Kubernetes pod log analysis to fix reactive stream error propagation boundaries?

Yes, Kubernetes pod log analysis identifies where reactive streams fail to enforce error propagation boundaries. By tracking repeating error patterns and silence windows, you can resolve the unbounded loops causing service outages.

What's the best way to resolve portal-wedging reactive storms in MeshWeaver portals?

Resolve portal-wedging reactive storms by performing root cause analysis on pod logs to distinguish deadlocks from exception loops. Enforce error propagation boundaries within the actor-model message hub to cure the reactive storm.