📦 deps(thirdparty): update snapshots
This commit is contained in:
ci[bot]
+1
-1
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "agentic-bundle-aas-observability-ir",
|
||||
"version": "14.2.0",
|
||||
"version": "14.3.1",
|
||||
"description": "Editorial \"AAS Observability IR\" bundle for Claude Code from Agentic Awesome Skills.",
|
||||
"author": {
|
||||
"name": "sickn33 and contributors",
|
||||
|
||||
+2
-2
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "aasb-aas-observability-ir",
|
||||
"version": "14.2.0",
|
||||
"version": "14.3.1",
|
||||
"description": "Install the \"AAS Observability IR\" workflow plugin from Agentic Awesome Skills.",
|
||||
"author": {
|
||||
"name": "sickn33 and contributors",
|
||||
@@ -20,7 +20,7 @@
|
||||
"interface": {
|
||||
"displayName": "AAS Observability IR",
|
||||
"shortDescription": "Design observability, SLOs, traces, dashboards, monitoring, incident response, troubleshooting, and postmortem workflows.",
|
||||
"longDescription": "Design observability, SLOs, traces, dashboards, monitoring, incident response, troubleshooting, and postmortem workflows. Operational work needs consistent procedure and proof gates, making it more plugin-worthy than isolated observability prompts. Recommended for: SRE teams, Backend teams owning production, Incident responders. Not for: Marketing campaign planning, Static document conversion. Covers Observability Engineer, Distributed Tracing, and 8 more skills.",
|
||||
"longDescription": "Design observability, SLOs, traces, dashboards, monitoring, incident response, troubleshooting, and postmortem workflows. Operational work needs consistent procedure and proof gates, making it more plugin-worthy than isolated observability prompts. Recommended for: SRE teams, Backend teams owning production, Incident responders. Not for: Marketing campaign planning, Static document conversion. Covers Observability Engineer, Observability And Instrumentation, and 8 more skills.",
|
||||
"developerName": "sickn33 and contributors",
|
||||
"category": "Specialized Product Plugins - Next Wave",
|
||||
"capabilities": [
|
||||
|
||||
-183
@@ -1,183 +0,0 @@
|
||||
---
|
||||
name: claude-monitor
|
||||
description: Monitor de performance do Claude Code e sistema local. Diagnostica lentidao, mede CPU/RAM/disco, verifica API latency e gera relatorios de saude do sistema.
|
||||
risk: safe
|
||||
source: community
|
||||
date_added: '2026-03-06'
|
||||
author: renat
|
||||
tags:
|
||||
- monitoring
|
||||
- performance
|
||||
- diagnostics
|
||||
- system-health
|
||||
tools:
|
||||
- claude-code
|
||||
- antigravity
|
||||
- cursor
|
||||
- gemini-cli
|
||||
- codex-cli
|
||||
---
|
||||
|
||||
# Claude Monitor — Diagnóstico de Performance
|
||||
|
||||
## Overview
|
||||
|
||||
Monitor de performance do Claude Code e sistema local. Diagnostica lentidao, mede CPU/RAM/disco, verifica API latency e gera relatorios de saude do sistema.
|
||||
|
||||
## When to Use This Skill
|
||||
|
||||
- When the user mentions "lento" or related topics
|
||||
- When the user mentions "lentidao" or related topics
|
||||
- When the user mentions "lag" or related topics
|
||||
- When the user mentions "lagado" or related topics
|
||||
- When the user mentions "travando" or related topics
|
||||
- When the user mentions "claude lento" or related topics
|
||||
|
||||
## Do Not Use This Skill When
|
||||
|
||||
- The task is unrelated to claude monitor
|
||||
- A simpler, more specific tool can handle the request
|
||||
- The user needs general-purpose assistance without domain expertise
|
||||
|
||||
## How It Works
|
||||
|
||||
Skill para diagnosticar e resolver problemas de lentidão no Claude Code e no sistema.
|
||||
Determina se o gargalo é local (PC) ou remoto (API Claude) e sugere ações corretivas.
|
||||
|
||||
## Quando Usar
|
||||
|
||||
- Usuário reclama que o Claude Code está lento ou travando
|
||||
- Troca de sessões de conversa demora para carregar
|
||||
- Respostas do Claude demoram muito
|
||||
- PC parece lento enquanto usa o Claude Code
|
||||
- Qualquer menção a performance, lag, lentidão
|
||||
|
||||
## 1. Diagnóstico Rápido (Health_Check.Py)
|
||||
|
||||
Rode SEMPRE como primeiro passo:
|
||||
|
||||
```bash
|
||||
python C:\Users\renat\skills\claude-monitor\scripts\health_check.py
|
||||
```
|
||||
|
||||
O script analisa em ~3 segundos:
|
||||
- **CPU**: Uso atual e por core. >80% = gargalo provável
|
||||
- **RAM**: Total, usada, disponível. >85% = pressão de memória
|
||||
- **Browsers**: Processos e RAM por browser. >5GB total = excesso de abas
|
||||
- **Claude Code**: Processos e RAM consumida
|
||||
- **Disco**: Espaço livre. <10% = impacto em swap/performance
|
||||
- **Rede**: Latência ao endpoint da API Claude
|
||||
- **Diagnóstico**: Classificação automática do problema com sugestões
|
||||
|
||||
## 2. Interpretar O Resultado
|
||||
|
||||
O script retorna um JSON com `diagnosis` contendo:
|
||||
|
||||
- `bottleneck`: "cpu" | "ram" | "browsers" | "disk" | "network" | "claude_api" | "ok"
|
||||
- `severity`: "critical" | "warning" | "ok"
|
||||
- `suggestions`: Lista de ações recomendadas
|
||||
- `summary`: Resumo em português para mostrar ao usuário
|
||||
|
||||
**Mostre o `summary` ao usuário** e ofereça executar as sugestões.
|
||||
|
||||
## 3. Ações Corretivas Automáticas
|
||||
|
||||
Baseado no diagnóstico, ofereça ao usuário:
|
||||
|
||||
#### Se CPU alta (>80%):
|
||||
- Listar processos consumindo mais CPU
|
||||
- Sugerir fechar processos pesados desnecessários
|
||||
- Verificar se Windows Update está rodando em background
|
||||
|
||||
#### Se browsers pesados (>5GB RAM ou >40 processos):
|
||||
```bash
|
||||
python C:\Users\renat\skills\claude-monitor\scripts\health_check.py --browsers-detail
|
||||
```
|
||||
Mostra RAM por browser e sugere quais fechar. **Nunca fechar processos sem permissão explícita do usuário.**
|
||||
|
||||
#### Se disco cheio (>85%):
|
||||
- Mostrar pastas maiores
|
||||
- Sugerir limpeza de Temp, cache de browsers, lixeira
|
||||
|
||||
#### Se rede lenta (latência >500ms):
|
||||
- Testar conexão com api.anthropic.com
|
||||
- Sugerir verificar VPN, proxy, ou conexão WiFi
|
||||
|
||||
## 4. Monitor Contínuo (Opcional)
|
||||
|
||||
Se o usuário quiser monitoramento em background:
|
||||
|
||||
```bash
|
||||
python C:\Users\renat\skills\claude-monitor\scripts\monitor.py --interval 30 --duration 300
|
||||
```
|
||||
|
||||
Parâmetros:
|
||||
- `--interval`: Segundos entre cada amostra (default: 30)
|
||||
- `--duration`: Duração total em segundos (default: 300 = 5 min)
|
||||
- `--output`: Caminho do arquivo de log (default: monitor_log.json)
|
||||
- `--alert-cpu`: Threshold de CPU para alerta (default: 80)
|
||||
- `--alert-ram`: Threshold de RAM % para alerta (default: 85)
|
||||
|
||||
O monitor salva snapshots periódicos e gera um relatório ao final com:
|
||||
- Picos de CPU e RAM
|
||||
- Tendência (melhorando/piorando/estável)
|
||||
- Eventos de alerta detectados
|
||||
- Recomendação final
|
||||
|
||||
## 5. Benchmark Da Api Claude (Opcional)
|
||||
|
||||
Para testar se a lentidão é da API:
|
||||
|
||||
```bash
|
||||
python C:\Users\renat\skills\claude-monitor\scripts\api_bench.py
|
||||
```
|
||||
|
||||
Mede o tempo de resposta do processo Claude Code local (não faz chamadas à API).
|
||||
Compara com tempos típicos e indica se está dentro do esperado.
|
||||
|
||||
## Thresholds De Referência
|
||||
|
||||
| Métrica | OK | Warning | Critical |
|
||||
|---------|-----|---------|----------|
|
||||
| CPU % | <60% | 60-85% | >85% |
|
||||
| RAM usada % | <70% | 70-85% | >85% |
|
||||
| RAM browsers | <3 GB | 3-6 GB | >6 GB |
|
||||
| Processos browser | <30 | 30-60 | >60 |
|
||||
| Disco livre | >15% | 10-15% | <10% |
|
||||
| Latência rede | <200ms | 200-500ms | >500ms |
|
||||
|
||||
## Dicas Para O Usuário
|
||||
|
||||
Quando apresentar o diagnóstico, inclua estas dicas contextuais:
|
||||
|
||||
- **Muitas abas = muito CPU/RAM**: Cada aba de browser é um processo separado.
|
||||
50 abas = 50 processos competindo por recursos.
|
||||
- **Claude Code é pesado**: Ele roda vários processos Electron. É normal consumir 3-5 GB.
|
||||
Mas se estiver usando >6 GB com várias sessões, considere fechar sessões antigas.
|
||||
- **Troca de sessão lenta**: Geralmente causada por CPU alta ou muitos processos competindo.
|
||||
A sessão precisa carregar o histórico da conversa, e se o CPU está ocupado, demora.
|
||||
- **Disco quase cheio**: Afeta a velocidade do swap (memória virtual) e pode causar
|
||||
lentidão generalizada.
|
||||
|
||||
## Dependências
|
||||
|
||||
- Python 3.10+
|
||||
- psutil (instalado automaticamente pelo script se não disponível)
|
||||
- Nenhuma API key necessária
|
||||
|
||||
## Best Practices
|
||||
|
||||
- Provide clear, specific context about your project and requirements
|
||||
- Review all suggestions before applying them to production code
|
||||
- Combine with other complementary skills for comprehensive analysis
|
||||
|
||||
## Common Pitfalls
|
||||
|
||||
- Using this skill for tasks outside its domain expertise
|
||||
- Applying recommendations without understanding your specific context
|
||||
- Not providing enough project context for accurate analysis
|
||||
|
||||
## Limitations
|
||||
- Use this skill only when the task clearly matches the scope described above.
|
||||
- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
|
||||
- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.
|
||||
-251
@@ -1,251 +0,0 @@
|
||||
#!/usr/bin/env python3
|
||||
"""
|
||||
Claude Monitor — Benchmark de Conectividade API
|
||||
|
||||
Testa latência e conectividade com a API do Claude.
|
||||
Não faz chamadas à API (não precisa de API key).
|
||||
Apenas verifica se a rede está funcionando e se o endpoint responde.
|
||||
|
||||
Uso:
|
||||
python api_bench.py # 5 testes de latência
|
||||
python api_bench.py --samples 10 # 10 testes
|
||||
python api_bench.py --json # Output JSON
|
||||
"""
|
||||
|
||||
import json
|
||||
import socket
|
||||
import ssl
|
||||
import subprocess
|
||||
import sys
|
||||
import time
|
||||
from datetime import datetime
|
||||
|
||||
try:
|
||||
import psutil
|
||||
except ImportError:
|
||||
subprocess.check_call([sys.executable, "-m", "pip", "install", "psutil", "--quiet"])
|
||||
import psutil
|
||||
|
||||
|
||||
ENDPOINTS = [
|
||||
{"name": "Claude API", "host": "api.anthropic.com", "port": 443},
|
||||
{"name": "Anthropic CDN", "host": "cdn.anthropic.com", "port": 443},
|
||||
{"name": "Google DNS", "host": "8.8.8.8", "port": 53},
|
||||
]
|
||||
|
||||
|
||||
def create_tls_context():
|
||||
"""Cria contexto TLS restringindo conexoes a TLS 1.2+."""
|
||||
context = ssl.create_default_context()
|
||||
if hasattr(ssl, "TLSVersion"):
|
||||
context.minimum_version = ssl.TLSVersion.TLSv1_2
|
||||
else:
|
||||
context.options |= getattr(ssl, "OP_NO_TLSv1", 0)
|
||||
context.options |= getattr(ssl, "OP_NO_TLSv1_1", 0)
|
||||
return context
|
||||
|
||||
|
||||
def test_tcp_latency(host, port, timeout=5):
|
||||
"""Testa latência TCP para um host:port."""
|
||||
try:
|
||||
start = time.time()
|
||||
sock = socket.create_connection((host, port), timeout=timeout)
|
||||
latency = (time.time() - start) * 1000 # ms
|
||||
sock.close()
|
||||
return {"reachable": True, "latency_ms": round(latency, 1)}
|
||||
except (socket.timeout, socket.error, OSError) as e:
|
||||
return {"reachable": False, "latency_ms": None, "error": str(e)}
|
||||
|
||||
|
||||
def test_tls_handshake(host, port=443, timeout=5):
|
||||
"""Testa tempo do handshake TLS."""
|
||||
try:
|
||||
context = create_tls_context()
|
||||
start = time.time()
|
||||
with socket.create_connection((host, port), timeout=timeout) as sock:
|
||||
with context.wrap_socket(sock, server_hostname=host) as ssock:
|
||||
handshake_time = (time.time() - start) * 1000
|
||||
return {
|
||||
"success": True,
|
||||
"handshake_ms": round(handshake_time, 1),
|
||||
"tls_version": ssock.version(),
|
||||
}
|
||||
except Exception as e:
|
||||
return {"success": False, "error": str(e)}
|
||||
|
||||
|
||||
def test_dns(hostname):
|
||||
"""Testa resolução DNS."""
|
||||
try:
|
||||
start = time.time()
|
||||
ip = socket.gethostbyname(hostname)
|
||||
dns_time = (time.time() - start) * 1000
|
||||
return {"resolved": True, "ip": ip, "dns_ms": round(dns_time, 1)}
|
||||
except socket.gaierror as e:
|
||||
return {"resolved": False, "error": str(e)}
|
||||
|
||||
|
||||
def check_network_interfaces():
|
||||
"""Verifica interfaces de rede ativas."""
|
||||
stats = psutil.net_if_stats()
|
||||
active = []
|
||||
for name, info in stats.items():
|
||||
if info.isup and info.speed > 0:
|
||||
active.append({
|
||||
"name": name,
|
||||
"speed_mbps": info.speed,
|
||||
"mtu": info.mtu,
|
||||
})
|
||||
return active
|
||||
|
||||
|
||||
def run_benchmark(samples=5):
|
||||
"""Roda o benchmark completo."""
|
||||
results = {
|
||||
"timestamp": datetime.now().isoformat(),
|
||||
"samples": samples,
|
||||
"endpoints": [],
|
||||
"dns": None,
|
||||
"tls": None,
|
||||
"network_interfaces": check_network_interfaces(),
|
||||
}
|
||||
|
||||
# DNS
|
||||
results["dns"] = test_dns("api.anthropic.com")
|
||||
|
||||
# TLS handshake
|
||||
results["tls"] = test_tls_handshake("api.anthropic.com")
|
||||
|
||||
# Latência por endpoint
|
||||
for ep in ENDPOINTS:
|
||||
latencies = []
|
||||
for _ in range(samples):
|
||||
result = test_tcp_latency(ep["host"], ep["port"])
|
||||
latencies.append(result)
|
||||
time.sleep(0.2)
|
||||
|
||||
valid = [r["latency_ms"] for r in latencies if r["reachable"] and r["latency_ms"]]
|
||||
|
||||
ep_result = {
|
||||
"name": ep["name"],
|
||||
"host": ep["host"],
|
||||
"port": ep["port"],
|
||||
"tests": latencies,
|
||||
}
|
||||
|
||||
if valid:
|
||||
ep_result["avg_ms"] = round(sum(valid) / len(valid), 1)
|
||||
ep_result["min_ms"] = round(min(valid), 1)
|
||||
ep_result["max_ms"] = round(max(valid), 1)
|
||||
ep_result["success_rate"] = round(len(valid) / samples * 100, 0)
|
||||
else:
|
||||
ep_result["avg_ms"] = None
|
||||
ep_result["success_rate"] = 0
|
||||
|
||||
results["endpoints"].append(ep_result)
|
||||
|
||||
# Diagnóstico
|
||||
api_ep = results["endpoints"][0]
|
||||
if api_ep.get("avg_ms") is None:
|
||||
results["diagnosis"] = {
|
||||
"status": "critical",
|
||||
"message": "API do Claude INACESSIVEL. Verifique sua conexao de internet.",
|
||||
}
|
||||
elif api_ep["avg_ms"] > 500:
|
||||
results["diagnosis"] = {
|
||||
"status": "warning",
|
||||
"message": (
|
||||
f"Latencia alta para API ({api_ep['avg_ms']}ms). "
|
||||
f"Conexao lenta pode causar atrasos no Claude Code."
|
||||
),
|
||||
}
|
||||
elif api_ep["avg_ms"] > 200:
|
||||
results["diagnosis"] = {
|
||||
"status": "ok",
|
||||
"message": (
|
||||
f"Latencia moderada ({api_ep['avg_ms']}ms). "
|
||||
f"Dentro do aceitavel mas pode ser melhor."
|
||||
),
|
||||
}
|
||||
else:
|
||||
results["diagnosis"] = {
|
||||
"status": "ok",
|
||||
"message": (
|
||||
f"Conexao excelente ({api_ep['avg_ms']}ms). "
|
||||
f"A rede NAO e o gargalo."
|
||||
),
|
||||
}
|
||||
|
||||
return results
|
||||
|
||||
|
||||
def format_results(results):
|
||||
"""Formata resultados para exibição."""
|
||||
lines = ["## Benchmark de Conectividade\n"]
|
||||
|
||||
# DNS
|
||||
dns = results["dns"]
|
||||
if dns.get("resolved"):
|
||||
lines.append(f"- DNS: api.anthropic.com -> {dns['ip']} ({dns['dns_ms']}ms)")
|
||||
else:
|
||||
lines.append(f"- DNS: FALHOU ({dns.get('error', 'desconhecido')})")
|
||||
|
||||
# TLS
|
||||
tls = results["tls"]
|
||||
if tls.get("success"):
|
||||
lines.append(f"- TLS: {tls['tls_version']} handshake em {tls['handshake_ms']}ms")
|
||||
else:
|
||||
lines.append(f"- TLS: FALHOU ({tls.get('error', 'desconhecido')})")
|
||||
|
||||
lines.append("")
|
||||
|
||||
# Endpoints
|
||||
lines.append("### Latencia por Endpoint")
|
||||
for ep in results["endpoints"]:
|
||||
if ep.get("avg_ms"):
|
||||
lines.append(
|
||||
f"- **{ep['name']}**: {ep['avg_ms']}ms avg "
|
||||
f"(min {ep['min_ms']}ms, max {ep['max_ms']}ms) "
|
||||
f"[{ep['success_rate']:.0f}% sucesso]"
|
||||
)
|
||||
else:
|
||||
lines.append(f"- **{ep['name']}**: INACESSIVEL")
|
||||
|
||||
# Interfaces
|
||||
lines.append("\n### Interfaces de Rede")
|
||||
for iface in results["network_interfaces"]:
|
||||
speed = iface["speed_mbps"]
|
||||
if speed >= 1000:
|
||||
speed_str = f"{speed/1000:.0f} Gbps"
|
||||
else:
|
||||
speed_str = f"{speed} Mbps"
|
||||
lines.append(f"- {iface['name']}: {speed_str}")
|
||||
|
||||
# Diagnóstico
|
||||
lines.append(f"\n### Diagnostico")
|
||||
diag = results["diagnosis"]
|
||||
status_map = {"critical": "[!!!]", "warning": "[!]", "ok": "[OK]"}
|
||||
lines.append(f"{status_map[diag['status']]} {diag['message']}")
|
||||
|
||||
return "\n".join(lines)
|
||||
|
||||
|
||||
def main():
|
||||
import argparse
|
||||
|
||||
parser = argparse.ArgumentParser(description="Claude Monitor - Benchmark de Conectividade")
|
||||
parser.add_argument("--samples", type=int, default=5, help="Numero de testes por endpoint")
|
||||
parser.add_argument("--json", action="store_true", help="Output JSON")
|
||||
args = parser.parse_args()
|
||||
|
||||
print(f"Testando conectividade ({args.samples} amostras por endpoint)...\n")
|
||||
results = run_benchmark(args.samples)
|
||||
|
||||
if args.json:
|
||||
print(json.dumps(results, indent=2, ensure_ascii=False))
|
||||
else:
|
||||
print(format_results(results))
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
-69
@@ -1,69 +0,0 @@
|
||||
"""
|
||||
Configurações e thresholds para o Claude Monitor.
|
||||
"""
|
||||
|
||||
# Thresholds de alerta
|
||||
THRESHOLDS = {
|
||||
"cpu": {
|
||||
"ok": 60,
|
||||
"warning": 85,
|
||||
# acima de warning = critical
|
||||
},
|
||||
"ram_percent": {
|
||||
"ok": 70,
|
||||
"warning": 85,
|
||||
},
|
||||
"browsers_ram_gb": {
|
||||
"ok": 3.0,
|
||||
"warning": 6.0,
|
||||
},
|
||||
"browsers_processes": {
|
||||
"ok": 30,
|
||||
"warning": 60,
|
||||
},
|
||||
"disk_free_percent": {
|
||||
"critical_below": 10,
|
||||
"warning_below": 15,
|
||||
},
|
||||
"network_latency_ms": {
|
||||
"ok": 200,
|
||||
"warning": 500,
|
||||
},
|
||||
}
|
||||
|
||||
# Nomes de processos de browser conhecidos
|
||||
BROWSER_NAMES = ["chrome", "msedge", "firefox", "brave", "opera", "vivaldi"]
|
||||
|
||||
# Nomes de processos do Claude Code
|
||||
CLAUDE_NAMES = ["claude"]
|
||||
|
||||
# Endpoint para teste de latência
|
||||
API_ENDPOINT = "api.anthropic.com"
|
||||
|
||||
# Monitor defaults
|
||||
MONITOR_DEFAULTS = {
|
||||
"interval": 30,
|
||||
"duration": 300,
|
||||
"alert_cpu": 80,
|
||||
"alert_ram": 85,
|
||||
}
|
||||
|
||||
|
||||
def classify(value, metric_name):
|
||||
"""Classifica um valor como 'ok', 'warning' ou 'critical'."""
|
||||
t = THRESHOLDS.get(metric_name, {})
|
||||
|
||||
# Métricas onde "abaixo" é ruim (disco livre)
|
||||
if "critical_below" in t:
|
||||
if value < t["critical_below"]:
|
||||
return "critical"
|
||||
elif value < t["warning_below"]:
|
||||
return "warning"
|
||||
return "ok"
|
||||
|
||||
# Métricas onde "acima" é ruim (CPU, RAM, latência)
|
||||
if value <= t.get("ok", 999999):
|
||||
return "ok"
|
||||
elif value <= t.get("warning", 999999):
|
||||
return "warning"
|
||||
return "critical"
|
||||
-362
@@ -1,362 +0,0 @@
|
||||
#!/usr/bin/env python3
|
||||
"""
|
||||
Claude Monitor — Diagnóstico Rápido de Performance
|
||||
|
||||
Analisa CPU, RAM, browsers, disco e rede em ~3 segundos.
|
||||
Identifica o gargalo principal e sugere ações corretivas.
|
||||
|
||||
Uso:
|
||||
python health_check.py # Diagnóstico completo
|
||||
python health_check.py --browsers-detail # Detalhe de browsers
|
||||
python health_check.py --json # Output JSON puro
|
||||
python health_check.py --quick # Só resumo (sem teste de rede)
|
||||
"""
|
||||
|
||||
import json
|
||||
import os
|
||||
import socket
|
||||
import subprocess
|
||||
import sys
|
||||
import time
|
||||
from datetime import datetime
|
||||
from pathlib import Path
|
||||
|
||||
# Garante que psutil está disponível
|
||||
try:
|
||||
import psutil
|
||||
except ImportError:
|
||||
print("Instalando psutil...")
|
||||
subprocess.check_call([sys.executable, "-m", "pip", "install", "psutil", "--quiet"])
|
||||
import psutil
|
||||
|
||||
# Importa config do mesmo diretório
|
||||
sys.path.insert(0, str(Path(__file__).parent))
|
||||
from config import (
|
||||
BROWSER_NAMES, CLAUDE_NAMES, API_ENDPOINT,
|
||||
THRESHOLDS, classify
|
||||
)
|
||||
|
||||
|
||||
def check_cpu():
|
||||
"""Verifica uso de CPU."""
|
||||
cpu_percent = psutil.cpu_percent(interval=1)
|
||||
cpu_count = psutil.cpu_count()
|
||||
per_cpu = psutil.cpu_percent(interval=0, percpu=True)
|
||||
|
||||
return {
|
||||
"percent": cpu_percent,
|
||||
"cores": cpu_count,
|
||||
"per_core": per_cpu,
|
||||
"status": classify(cpu_percent, "cpu"),
|
||||
}
|
||||
|
||||
|
||||
def check_ram():
|
||||
"""Verifica uso de RAM."""
|
||||
ram = psutil.virtual_memory()
|
||||
swap = psutil.swap_memory()
|
||||
|
||||
return {
|
||||
"total_gb": round(ram.total / 1024**3, 1),
|
||||
"used_gb": round(ram.used / 1024**3, 1),
|
||||
"available_gb": round(ram.available / 1024**3, 1),
|
||||
"percent": ram.percent,
|
||||
"swap_used_gb": round(swap.used / 1024**3, 1),
|
||||
"swap_percent": swap.percent,
|
||||
"status": classify(ram.percent, "ram_percent"),
|
||||
}
|
||||
|
||||
|
||||
def check_browsers(detail=False):
|
||||
"""Verifica processos de browser e consumo de RAM."""
|
||||
browsers = {}
|
||||
all_procs = []
|
||||
|
||||
for proc in psutil.process_iter(["pid", "name", "memory_info"]):
|
||||
try:
|
||||
info = proc.info
|
||||
name_lower = info["name"].lower()
|
||||
ram_mb = info["memory_info"].rss / 1024**2
|
||||
|
||||
for bname in BROWSER_NAMES:
|
||||
if bname in name_lower:
|
||||
if bname not in browsers:
|
||||
browsers[bname] = {"count": 0, "ram_mb": 0, "pids": []}
|
||||
browsers[bname]["count"] += 1
|
||||
browsers[bname]["ram_mb"] += ram_mb
|
||||
if detail:
|
||||
browsers[bname]["pids"].append({
|
||||
"pid": info["pid"],
|
||||
"ram_mb": round(ram_mb, 0)
|
||||
})
|
||||
break
|
||||
except (psutil.NoSuchProcess, psutil.AccessDenied):
|
||||
pass
|
||||
|
||||
total_ram_gb = sum(b["ram_mb"] for b in browsers.values()) / 1024
|
||||
total_procs = sum(b["count"] for b in browsers.values())
|
||||
|
||||
# Formata para output
|
||||
for bname in browsers:
|
||||
browsers[bname]["ram_mb"] = round(browsers[bname]["ram_mb"], 0)
|
||||
|
||||
return {
|
||||
"browsers": browsers,
|
||||
"total_ram_gb": round(total_ram_gb, 1),
|
||||
"total_processes": total_procs,
|
||||
"ram_status": classify(total_ram_gb, "browsers_ram_gb"),
|
||||
"process_status": classify(total_procs, "browsers_processes"),
|
||||
}
|
||||
|
||||
|
||||
def check_claude_processes():
|
||||
"""Verifica processos do Claude Code."""
|
||||
claude_procs = []
|
||||
total_ram = 0
|
||||
|
||||
for proc in psutil.process_iter(["pid", "name", "memory_info", "cpu_percent"]):
|
||||
try:
|
||||
info = proc.info
|
||||
name_lower = info["name"].lower()
|
||||
|
||||
for cname in CLAUDE_NAMES:
|
||||
if cname in name_lower:
|
||||
ram_mb = info["memory_info"].rss / 1024**2
|
||||
claude_procs.append({
|
||||
"pid": info["pid"],
|
||||
"name": info["name"],
|
||||
"ram_mb": round(ram_mb, 0),
|
||||
})
|
||||
total_ram += ram_mb
|
||||
break
|
||||
except (psutil.NoSuchProcess, psutil.AccessDenied):
|
||||
pass
|
||||
|
||||
claude_procs.sort(key=lambda x: x["ram_mb"], reverse=True)
|
||||
|
||||
return {
|
||||
"count": len(claude_procs),
|
||||
"total_ram_gb": round(total_ram / 1024, 1),
|
||||
"processes": claude_procs[:10], # Top 10
|
||||
}
|
||||
|
||||
|
||||
def check_disk():
|
||||
"""Verifica espaço em disco."""
|
||||
disk = psutil.disk_usage("C:/")
|
||||
free_percent = 100 - disk.percent
|
||||
|
||||
return {
|
||||
"total_gb": round(disk.total / 1024**3, 0),
|
||||
"used_gb": round(disk.used / 1024**3, 0),
|
||||
"free_gb": round(disk.free / 1024**3, 0),
|
||||
"used_percent": disk.percent,
|
||||
"free_percent": round(free_percent, 1),
|
||||
"status": classify(free_percent, "disk_free_percent"),
|
||||
}
|
||||
|
||||
|
||||
def check_network():
|
||||
"""Testa latência até a API do Claude."""
|
||||
try:
|
||||
start = time.time()
|
||||
sock = socket.create_connection((API_ENDPOINT, 443), timeout=5)
|
||||
latency_ms = round((time.time() - start) * 1000, 0)
|
||||
sock.close()
|
||||
|
||||
return {
|
||||
"latency_ms": latency_ms,
|
||||
"endpoint": API_ENDPOINT,
|
||||
"reachable": True,
|
||||
"status": classify(latency_ms, "network_latency_ms"),
|
||||
}
|
||||
except (socket.timeout, socket.error, OSError) as e:
|
||||
return {
|
||||
"latency_ms": None,
|
||||
"endpoint": API_ENDPOINT,
|
||||
"reachable": False,
|
||||
"status": "critical",
|
||||
"error": str(e),
|
||||
}
|
||||
|
||||
|
||||
def check_top_processes(n=10):
|
||||
"""Lista os N processos que mais consomem RAM."""
|
||||
procs = []
|
||||
for proc in psutil.process_iter(["pid", "name", "memory_info"]):
|
||||
try:
|
||||
info = proc.info
|
||||
procs.append({
|
||||
"name": info["name"],
|
||||
"ram_mb": round(info["memory_info"].rss / 1024**2, 0),
|
||||
"pid": info["pid"],
|
||||
})
|
||||
except (psutil.NoSuchProcess, psutil.AccessDenied):
|
||||
pass
|
||||
|
||||
procs.sort(key=lambda x: x["ram_mb"], reverse=True)
|
||||
return procs[:n]
|
||||
|
||||
|
||||
def diagnose(results):
|
||||
"""Analisa os resultados e gera diagnóstico."""
|
||||
issues = []
|
||||
suggestions = []
|
||||
bottleneck = "ok"
|
||||
severity = "ok"
|
||||
|
||||
cpu = results["cpu"]
|
||||
ram = results["ram"]
|
||||
browsers = results["browsers"]
|
||||
disk = results["disk"]
|
||||
network = results.get("network", {})
|
||||
claude = results["claude"]
|
||||
|
||||
# CPU
|
||||
if cpu["status"] == "critical":
|
||||
issues.append(f"CPU a {cpu['percent']}% (CRITICO)")
|
||||
suggestions.append("Fechar aplicativos pesados ou abas de browser desnecessarias")
|
||||
suggestions.append("Verificar se Windows Update ou antivirus esta rodando em background")
|
||||
bottleneck = "cpu"
|
||||
severity = "critical"
|
||||
elif cpu["status"] == "warning":
|
||||
issues.append(f"CPU a {cpu['percent']}% (elevada)")
|
||||
suggestions.append("Considerar fechar algumas abas de browser")
|
||||
if severity != "critical":
|
||||
bottleneck = "cpu"
|
||||
severity = "warning"
|
||||
|
||||
# RAM
|
||||
if ram["status"] == "critical":
|
||||
issues.append(f"RAM a {ram['percent']}% ({ram['used_gb']} de {ram['total_gb']} GB)")
|
||||
suggestions.append("Fechar browsers ou aplicativos para liberar memoria")
|
||||
if severity != "critical":
|
||||
bottleneck = "ram"
|
||||
severity = "critical"
|
||||
elif ram["status"] == "warning":
|
||||
issues.append(f"RAM a {ram['percent']}% (monitorar)")
|
||||
|
||||
# Browsers
|
||||
if browsers["ram_status"] == "critical":
|
||||
issues.append(f"Browsers consumindo {browsers['total_ram_gb']} GB ({browsers['total_processes']} processos)")
|
||||
suggestions.append("Fechar abas desnecessarias nos browsers")
|
||||
browser_detail = []
|
||||
for bname, info in browsers["browsers"].items():
|
||||
browser_detail.append(f" - {bname}: {info['count']} processos, {info['ram_mb']:.0f} MB")
|
||||
suggestions.append("Detalhamento:\n" + "\n".join(browser_detail))
|
||||
if bottleneck == "ok":
|
||||
bottleneck = "browsers"
|
||||
if severity == "ok":
|
||||
severity = "warning"
|
||||
elif browsers["ram_status"] == "warning":
|
||||
issues.append(f"Browsers usando {browsers['total_ram_gb']} GB (moderado)")
|
||||
|
||||
# Disco
|
||||
if disk["status"] == "critical":
|
||||
issues.append(f"Disco quase cheio: apenas {disk['free_gb']:.0f} GB livres ({disk['free_percent']}%)")
|
||||
suggestions.append("Limpar arquivos temporarios, cache e lixeira")
|
||||
suggestions.append("Verificar pasta Downloads e Temp por arquivos grandes")
|
||||
if bottleneck == "ok":
|
||||
bottleneck = "disk"
|
||||
severity = "warning"
|
||||
elif disk["status"] == "warning":
|
||||
issues.append(f"Disco com {disk['free_gb']:.0f} GB livres ({disk['free_percent']}%)")
|
||||
|
||||
# Rede
|
||||
if network.get("status") == "critical":
|
||||
if not network.get("reachable"):
|
||||
issues.append("API do Claude INACESSIVEL")
|
||||
suggestions.append("Verificar conexao com internet")
|
||||
suggestions.append("Verificar se VPN ou proxy esta bloqueando")
|
||||
bottleneck = "network"
|
||||
severity = "critical"
|
||||
else:
|
||||
issues.append(f"Latencia alta para API: {network['latency_ms']}ms")
|
||||
suggestions.append("Verificar qualidade da conexao WiFi/cabo")
|
||||
if bottleneck == "ok":
|
||||
bottleneck = "network"
|
||||
severity = "warning"
|
||||
|
||||
# Claude Code RAM
|
||||
if claude["total_ram_gb"] > 8:
|
||||
issues.append(f"Claude Code usando {claude['total_ram_gb']} GB ({claude['count']} processos)")
|
||||
suggestions.append("Considerar fechar sessoes de conversa antigas no Claude Code")
|
||||
|
||||
# Tudo ok
|
||||
if not issues:
|
||||
issues.append("Sistema saudavel, sem gargalos detectados")
|
||||
suggestions.append("A lentidao pode ser temporaria (pico na API do Claude)")
|
||||
suggestions.append("Tente trocar de sessao novamente em alguns segundos")
|
||||
|
||||
# Gerar resumo em PT-BR
|
||||
summary_lines = ["## Diagnostico de Performance\n"]
|
||||
|
||||
status_emoji = {"critical": "[!!!]", "warning": "[!]", "ok": "[OK]"}
|
||||
summary_lines.append(f"**Status geral: {status_emoji[severity]} {severity.upper()}**\n")
|
||||
|
||||
if bottleneck != "ok":
|
||||
summary_lines.append(f"**Gargalo principal: {bottleneck.upper()}**\n")
|
||||
|
||||
summary_lines.append("### Problemas detectados:")
|
||||
for issue in issues:
|
||||
summary_lines.append(f"- {issue}")
|
||||
|
||||
summary_lines.append("\n### Acoes recomendadas:")
|
||||
for i, sug in enumerate(suggestions, 1):
|
||||
if "\n" in sug:
|
||||
summary_lines.append(f"{i}. {sug}")
|
||||
else:
|
||||
summary_lines.append(f"{i}. {sug}")
|
||||
|
||||
summary_lines.append(f"\n### Numeros-chave:")
|
||||
summary_lines.append(f"- CPU: {cpu['percent']}% | RAM: {ram['percent']}% ({ram['used_gb']}/{ram['total_gb']} GB)")
|
||||
summary_lines.append(f"- Browsers: {browsers['total_processes']} processos, {browsers['total_ram_gb']} GB")
|
||||
summary_lines.append(f"- Claude Code: {claude['count']} processos, {claude['total_ram_gb']} GB")
|
||||
summary_lines.append(f"- Disco C: {disk['free_gb']:.0f} GB livres ({disk['free_percent']}%)")
|
||||
if network.get("latency_ms"):
|
||||
summary_lines.append(f"- Latencia API: {network['latency_ms']}ms")
|
||||
|
||||
return {
|
||||
"bottleneck": bottleneck,
|
||||
"severity": severity,
|
||||
"issues": issues,
|
||||
"suggestions": suggestions,
|
||||
"summary": "\n".join(summary_lines),
|
||||
}
|
||||
|
||||
|
||||
def main():
|
||||
import argparse
|
||||
|
||||
parser = argparse.ArgumentParser(description="Claude Monitor - Diagnostico Rapido")
|
||||
parser.add_argument("--browsers-detail", action="store_true", help="Mostra detalhes por browser")
|
||||
parser.add_argument("--json", action="store_true", help="Output em JSON puro")
|
||||
parser.add_argument("--quick", action="store_true", help="Pula teste de rede")
|
||||
args = parser.parse_args()
|
||||
|
||||
results = {}
|
||||
|
||||
# Coleta dados
|
||||
results["timestamp"] = datetime.now().isoformat()
|
||||
results["cpu"] = check_cpu()
|
||||
results["ram"] = check_ram()
|
||||
results["browsers"] = check_browsers(detail=args.browsers_detail)
|
||||
results["claude"] = check_claude_processes()
|
||||
results["disk"] = check_disk()
|
||||
results["top_processes"] = check_top_processes(15)
|
||||
|
||||
if not args.quick:
|
||||
results["network"] = check_network()
|
||||
|
||||
# Diagnóstico
|
||||
results["diagnosis"] = diagnose(results)
|
||||
|
||||
if args.json:
|
||||
print(json.dumps(results, indent=2, ensure_ascii=False))
|
||||
else:
|
||||
print(results["diagnosis"]["summary"])
|
||||
print(f"\n(Para output completo em JSON, use: python health_check.py --json)")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
-309
@@ -1,309 +0,0 @@
|
||||
#!/usr/bin/env python3
|
||||
"""
|
||||
Claude Monitor — Monitor Contínuo de Performance
|
||||
|
||||
Coleta snapshots periódicos de CPU, RAM e browsers.
|
||||
Gera relatório com tendências e alertas ao final.
|
||||
|
||||
Uso:
|
||||
python monitor.py # 5 min, amostras a cada 30s
|
||||
python monitor.py --interval 10 --duration 120 # 2 min, amostras a cada 10s
|
||||
python monitor.py --output meu_log.json # Salvar em arquivo específico
|
||||
"""
|
||||
|
||||
import json
|
||||
import os
|
||||
import signal
|
||||
import subprocess
|
||||
import sys
|
||||
import time
|
||||
from datetime import datetime
|
||||
from pathlib import Path
|
||||
|
||||
|
||||
def safe_user_path(path_value, base_dir="."):
|
||||
"""Resolve a CLI path under the current workspace."""
|
||||
if base_dir != ".":
|
||||
raise ValueError("Custom base directories are not supported for CLI paths")
|
||||
base_path = Path.cwd().resolve()
|
||||
resolved_path = Path(path_value).expanduser().resolve()
|
||||
try:
|
||||
resolved_path.relative_to(base_path)
|
||||
except ValueError as exc:
|
||||
raise ValueError(f"Path escapes allowed directory: {path_value}") from exc
|
||||
return resolved_path
|
||||
|
||||
try:
|
||||
import psutil
|
||||
except ImportError:
|
||||
subprocess.check_call([sys.executable, "-m", "pip", "install", "psutil", "--quiet"])
|
||||
import psutil
|
||||
|
||||
sys.path.insert(0, str(Path(__file__).parent))
|
||||
from config import BROWSER_NAMES, CLAUDE_NAMES, MONITOR_DEFAULTS
|
||||
|
||||
|
||||
def take_snapshot():
|
||||
"""Coleta um snapshot rápido do sistema."""
|
||||
cpu = psutil.cpu_percent(interval=0.5)
|
||||
ram = psutil.virtual_memory()
|
||||
|
||||
# Browser totals
|
||||
browser_ram = 0
|
||||
browser_count = 0
|
||||
for proc in psutil.process_iter(["name", "memory_info"]):
|
||||
try:
|
||||
name = proc.info["name"].lower()
|
||||
for bname in BROWSER_NAMES:
|
||||
if bname in name:
|
||||
browser_ram += proc.info["memory_info"].rss
|
||||
browser_count += 1
|
||||
break
|
||||
except (psutil.NoSuchProcess, psutil.AccessDenied):
|
||||
pass
|
||||
|
||||
# Claude totals
|
||||
claude_ram = 0
|
||||
claude_count = 0
|
||||
for proc in psutil.process_iter(["name", "memory_info"]):
|
||||
try:
|
||||
name = proc.info["name"].lower()
|
||||
for cname in CLAUDE_NAMES:
|
||||
if cname in name:
|
||||
claude_ram += proc.info["memory_info"].rss
|
||||
claude_count += 1
|
||||
break
|
||||
except (psutil.NoSuchProcess, psutil.AccessDenied):
|
||||
pass
|
||||
|
||||
return {
|
||||
"timestamp": datetime.now().isoformat(),
|
||||
"cpu_percent": cpu,
|
||||
"ram_percent": ram.percent,
|
||||
"ram_used_gb": round(ram.used / 1024**3, 2),
|
||||
"ram_available_gb": round(ram.available / 1024**3, 2),
|
||||
"browser_ram_gb": round(browser_ram / 1024**3, 2),
|
||||
"browser_processes": browser_count,
|
||||
"claude_ram_gb": round(claude_ram / 1024**3, 2),
|
||||
"claude_processes": claude_count,
|
||||
}
|
||||
|
||||
|
||||
def analyze_snapshots(snapshots, alert_cpu, alert_ram):
|
||||
"""Analisa os snapshots coletados e gera relatório."""
|
||||
if not snapshots:
|
||||
return {"error": "Nenhum snapshot coletado"}
|
||||
|
||||
n = len(snapshots)
|
||||
cpu_values = [s["cpu_percent"] for s in snapshots]
|
||||
ram_values = [s["ram_percent"] for s in snapshots]
|
||||
browser_ram_values = [s["browser_ram_gb"] for s in snapshots]
|
||||
|
||||
# Alertas
|
||||
alerts = []
|
||||
for s in snapshots:
|
||||
if s["cpu_percent"] >= alert_cpu:
|
||||
alerts.append({
|
||||
"time": s["timestamp"],
|
||||
"type": "cpu",
|
||||
"value": s["cpu_percent"],
|
||||
"threshold": alert_cpu,
|
||||
})
|
||||
if s["ram_percent"] >= alert_ram:
|
||||
alerts.append({
|
||||
"time": s["timestamp"],
|
||||
"type": "ram",
|
||||
"value": s["ram_percent"],
|
||||
"threshold": alert_ram,
|
||||
})
|
||||
|
||||
# Tendência (compara primeira metade com segunda metade)
|
||||
mid = n // 2
|
||||
if mid > 0:
|
||||
cpu_first = sum(cpu_values[:mid]) / mid
|
||||
cpu_second = sum(cpu_values[mid:]) / (n - mid)
|
||||
ram_first = sum(ram_values[:mid]) / mid
|
||||
ram_second = sum(ram_values[mid:]) / (n - mid)
|
||||
|
||||
cpu_diff = cpu_second - cpu_first
|
||||
ram_diff = ram_second - ram_first
|
||||
|
||||
if abs(cpu_diff) < 5 and abs(ram_diff) < 3:
|
||||
trend = "estavel"
|
||||
elif cpu_diff > 5 or ram_diff > 3:
|
||||
trend = "piorando"
|
||||
else:
|
||||
trend = "melhorando"
|
||||
else:
|
||||
trend = "insuficiente"
|
||||
cpu_diff = 0
|
||||
ram_diff = 0
|
||||
|
||||
# Resumo
|
||||
report = {
|
||||
"samples": n,
|
||||
"duration_seconds": round(
|
||||
(datetime.fromisoformat(snapshots[-1]["timestamp"]) -
|
||||
datetime.fromisoformat(snapshots[0]["timestamp"])).total_seconds(), 0
|
||||
) if n > 1 else 0,
|
||||
"cpu": {
|
||||
"avg": round(sum(cpu_values) / n, 1),
|
||||
"max": round(max(cpu_values), 1),
|
||||
"min": round(min(cpu_values), 1),
|
||||
},
|
||||
"ram": {
|
||||
"avg_percent": round(sum(ram_values) / n, 1),
|
||||
"max_percent": round(max(ram_values), 1),
|
||||
"avg_used_gb": round(sum(s["ram_used_gb"] for s in snapshots) / n, 1),
|
||||
},
|
||||
"browsers": {
|
||||
"avg_ram_gb": round(sum(browser_ram_values) / n, 1),
|
||||
"max_ram_gb": round(max(browser_ram_values), 1),
|
||||
"avg_processes": round(sum(s["browser_processes"] for s in snapshots) / n, 0),
|
||||
},
|
||||
"trend": trend,
|
||||
"trend_detail": {
|
||||
"cpu_change": round(cpu_diff, 1),
|
||||
"ram_change": round(ram_diff, 1),
|
||||
},
|
||||
"alerts_count": len(alerts),
|
||||
"alerts": alerts[:20], # Máximo 20 alertas no relatório
|
||||
}
|
||||
|
||||
# Recomendação final
|
||||
if report["cpu"]["avg"] > alert_cpu:
|
||||
report["recommendation"] = (
|
||||
f"CPU consistentemente alta (media {report['cpu']['avg']}%). "
|
||||
f"Fechar aplicativos pesados e abas de browser desnecessarias."
|
||||
)
|
||||
elif len(alerts) > n * 0.3:
|
||||
report["recommendation"] = (
|
||||
f"Alertas frequentes ({len(alerts)} de {n} amostras). "
|
||||
f"Sistema sob pressao intermitente. Reduzir carga."
|
||||
)
|
||||
elif trend == "piorando":
|
||||
report["recommendation"] = (
|
||||
f"Tendencia de piora detectada (CPU {'+' if cpu_diff > 0 else ''}{cpu_diff:.0f}%, "
|
||||
f"RAM {'+' if ram_diff > 0 else ''}{ram_diff:.0f}%). Monitorar."
|
||||
)
|
||||
else:
|
||||
report["recommendation"] = "Sistema estavel durante o monitoramento."
|
||||
|
||||
return report
|
||||
|
||||
|
||||
def format_report(report):
|
||||
"""Formata o relatório para exibição."""
|
||||
lines = ["## Relatorio de Monitoramento\n"]
|
||||
lines.append(f"- **Amostras**: {report['samples']} em {report['duration_seconds']}s")
|
||||
lines.append(f"- **Tendencia**: {report['trend'].upper()}")
|
||||
lines.append(f"- **Alertas**: {report['alerts_count']}\n")
|
||||
|
||||
lines.append("### CPU")
|
||||
lines.append(f"- Media: {report['cpu']['avg']}%")
|
||||
lines.append(f"- Max: {report['cpu']['max']}% | Min: {report['cpu']['min']}%\n")
|
||||
|
||||
lines.append("### RAM")
|
||||
lines.append(f"- Media: {report['ram']['avg_percent']}% ({report['ram']['avg_used_gb']} GB)")
|
||||
lines.append(f"- Pico: {report['ram']['max_percent']}%\n")
|
||||
|
||||
lines.append("### Browsers")
|
||||
lines.append(f"- Media RAM: {report['browsers']['avg_ram_gb']} GB")
|
||||
lines.append(f"- Pico RAM: {report['browsers']['max_ram_gb']} GB")
|
||||
lines.append(f"- Media processos: {report['browsers']['avg_processes']}\n")
|
||||
|
||||
lines.append(f"### Recomendacao")
|
||||
lines.append(f"{report['recommendation']}")
|
||||
|
||||
return "\n".join(lines)
|
||||
|
||||
|
||||
def main():
|
||||
import argparse
|
||||
|
||||
parser = argparse.ArgumentParser(description="Claude Monitor - Monitor Continuo")
|
||||
parser.add_argument("--interval", type=int, default=MONITOR_DEFAULTS["interval"],
|
||||
help=f"Segundos entre amostras (default: {MONITOR_DEFAULTS['interval']})")
|
||||
parser.add_argument("--duration", type=int, default=MONITOR_DEFAULTS["duration"],
|
||||
help=f"Duracao total em segundos (default: {MONITOR_DEFAULTS['duration']})")
|
||||
parser.add_argument("--output", type=str, default=None,
|
||||
help="Arquivo de saida JSON")
|
||||
parser.add_argument("--alert-cpu", type=int, default=MONITOR_DEFAULTS["alert_cpu"],
|
||||
help=f"Threshold CPU para alerta (default: {MONITOR_DEFAULTS['alert_cpu']})")
|
||||
parser.add_argument("--alert-ram", type=int, default=MONITOR_DEFAULTS["alert_ram"],
|
||||
help=f"Threshold RAM para alerta (default: {MONITOR_DEFAULTS['alert_ram']})")
|
||||
parser.add_argument("--json", action="store_true", help="Output em JSON")
|
||||
args = parser.parse_args()
|
||||
|
||||
snapshots = []
|
||||
start_time = time.time()
|
||||
sample_count = 0
|
||||
expected_samples = args.duration // args.interval
|
||||
|
||||
print(f"Monitorando por {args.duration}s (amostra a cada {args.interval}s)...")
|
||||
print(f"Esperando {expected_samples} amostras. Ctrl+C para parar.\n")
|
||||
|
||||
# Permite interromper com Ctrl+C
|
||||
interrupted = False
|
||||
|
||||
def handle_interrupt(sig, frame):
|
||||
nonlocal interrupted
|
||||
interrupted = True
|
||||
print("\nInterrompido pelo usuario. Gerando relatorio...\n")
|
||||
|
||||
signal.signal(signal.SIGINT, handle_interrupt)
|
||||
|
||||
while not interrupted and (time.time() - start_time) < args.duration:
|
||||
snapshot = take_snapshot()
|
||||
snapshots.append(snapshot)
|
||||
sample_count += 1
|
||||
|
||||
# Print inline progress
|
||||
print(
|
||||
f"[{sample_count}/{expected_samples}] "
|
||||
f"CPU: {snapshot['cpu_percent']:5.1f}% | "
|
||||
f"RAM: {snapshot['ram_percent']:5.1f}% | "
|
||||
f"Browsers: {snapshot['browser_ram_gb']:.1f}GB ({snapshot['browser_processes']} proc) | "
|
||||
f"Claude: {snapshot['claude_ram_gb']:.1f}GB ({snapshot['claude_processes']} proc)"
|
||||
)
|
||||
|
||||
# Espera até a próxima amostra
|
||||
elapsed = time.time() - start_time
|
||||
next_sample_at = sample_count * args.interval
|
||||
sleep_time = max(0, next_sample_at - elapsed)
|
||||
if sleep_time > 0 and not interrupted:
|
||||
time.sleep(sleep_time)
|
||||
|
||||
# Analisa
|
||||
report = analyze_snapshots(snapshots, args.alert_cpu, args.alert_ram)
|
||||
|
||||
# Salva log
|
||||
output_data = {
|
||||
"config": {
|
||||
"interval": args.interval,
|
||||
"duration": args.duration,
|
||||
"alert_cpu": args.alert_cpu,
|
||||
"alert_ram": args.alert_ram,
|
||||
},
|
||||
"snapshots": snapshots,
|
||||
"report": report,
|
||||
}
|
||||
|
||||
if args.output:
|
||||
output_path = args.output
|
||||
else:
|
||||
output_path = f"monitor_log_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json"
|
||||
|
||||
with safe_user_path(output_path).open("w", encoding="utf-8") as f:
|
||||
f.write(json.dumps(output_data, indent=2, ensure_ascii=False))
|
||||
|
||||
print(f"\nLog salvo em: {output_path}\n")
|
||||
|
||||
if args.json:
|
||||
print(json.dumps(report, indent=2, ensure_ascii=False))
|
||||
else:
|
||||
print(format_report(report))
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
+216
@@ -0,0 +1,216 @@
|
||||
---
|
||||
name: observability-and-instrumentation
|
||||
description: Instruments code so production behavior is visible and diagnosable. Use when adding logging, metrics, tracing, or alerting. Use when shipping any feature that runs in production and you need evidence it works. Use when production issues are reported but you can't tell what happened...
|
||||
risk: unknown
|
||||
source: https://github.com/addyosmani/agent-skills/tree/main/skills/observability-and-instrumentation
|
||||
source_repo: addyosmani/agent-skills
|
||||
source_type: community
|
||||
date_added: 2026-07-01
|
||||
license: MIT
|
||||
license_source: https://github.com/addyosmani/agent-skills/blob/main/LICENSE
|
||||
---
|
||||
|
||||
# Observability and Instrumentation
|
||||
|
||||
## Overview
|
||||
|
||||
Code you can't observe is code you can't operate. Observability is the ability to answer "what is the system doing and why?" from the outside, using the telemetry the code emits. Instrumentation is not a post-launch add-on — it's written alongside the feature, the same way tests are. If a feature ships without telemetry, the first user-reported bug becomes archaeology instead of a query.
|
||||
|
||||
## When to Use
|
||||
|
||||
- Building any feature that will run in production
|
||||
- Adding a new service, endpoint, background job, or external integration
|
||||
- A production incident took too long to diagnose ("we couldn't tell what happened")
|
||||
- Setting up or reviewing alerting rules
|
||||
- Reviewing a PR that adds I/O, retries, queues, or cross-service calls
|
||||
|
||||
**NOT for:**
|
||||
- Diagnosing a failure happening right now — use the `debugging-and-error-recovery` skill (observability is what makes that skill fast next time)
|
||||
- Profiling and optimizing measured slowness — use the `performance-optimization` skill
|
||||
- Launch-day monitoring checklists and rollback triggers — see the `shipping-and-launch` skill; this skill covers the instrumentation that feeds them
|
||||
|
||||
## Process
|
||||
|
||||
### 1. Define "working" before instrumenting
|
||||
|
||||
Telemetry without a question is noise. Before adding any instrumentation, write down 2–4 questions an on-call engineer will ask about this feature:
|
||||
|
||||
```
|
||||
FEATURE: checkout payment retry
|
||||
QUESTIONS ON-CALL WILL ASK:
|
||||
1. What fraction of payments succeed on first attempt vs after retry?
|
||||
2. When a payment fails permanently, why? (provider error? timeout? validation?)
|
||||
3. Is the payment provider slower than usual?
|
||||
→ Every signal below must help answer one of these.
|
||||
```
|
||||
|
||||
If you can't name the questions, you're not ready to instrument — you'll log everything and learn nothing.
|
||||
|
||||
### 2. Pick the right signal for each question
|
||||
|
||||
| Signal | Answers | Cost profile | Example |
|
||||
|---|---|---|---|
|
||||
| **Structured log** | "What happened in this specific case?" | Per-event; grows with traffic | `payment_failed` with provider error code |
|
||||
| **Metric** | "How often / how fast, in aggregate?" | Fixed per series; cheap to query | p99 latency of provider calls |
|
||||
| **Trace** | "Where did time go across services?" | Per-request; usually sampled | One slow checkout, broken down by hop |
|
||||
|
||||
Rule of thumb: metrics tell you **that** something is wrong, traces tell you **where**, logs tell you **why**.
|
||||
|
||||
### 3. Structured logging
|
||||
|
||||
Log events, not prose. Every log line is a JSON object with a stable event name and machine-readable fields:
|
||||
|
||||
```typescript
|
||||
// BAD: string interpolation — unqueryable, inconsistent
|
||||
logger.info(`Payment ${id} failed for user ${userId} after ${n} retries`);
|
||||
|
||||
// GOOD: stable event name + structured fields
|
||||
logger.warn({
|
||||
event: 'payment_failed',
|
||||
paymentId: id,
|
||||
provider: 'stripe',
|
||||
errorCode: err.code,
|
||||
attempt: n,
|
||||
}, 'payment failed');
|
||||
```
|
||||
|
||||
**Log levels — use them consistently:**
|
||||
|
||||
| Level | Meaning | On-call action |
|
||||
|---|---|---|
|
||||
| `error` | Invariant broken; someone may need to act | Investigate |
|
||||
| `warn` | Degraded but handled (retry succeeded, fallback used) | Watch for trends |
|
||||
| `info` | Significant business event (order placed, job finished) | None |
|
||||
| `debug` | Diagnostic detail | Off in production by default |
|
||||
|
||||
**Correlation IDs are mandatory.** Generate (or accept) a request ID at the system boundary and attach it to every log line, span, and outbound call. Without it, you cannot reconstruct a single request from interleaved logs:
|
||||
|
||||
```typescript
|
||||
// Express: child logger per request, ID propagated downstream
|
||||
app.use((req, res, next) => {
|
||||
req.id = req.headers['x-request-id'] ?? crypto.randomUUID();
|
||||
req.log = logger.child({ requestId: req.id });
|
||||
res.setHeader('x-request-id', req.id);
|
||||
next();
|
||||
});
|
||||
```
|
||||
|
||||
**Never log secrets, tokens, passwords, or full PII.** This is a hard rule from the `security-and-hardening` skill — telemetry pipelines are a classic data-leak path. Allowlist fields; don't log whole request bodies.
|
||||
|
||||
### 4. Metrics
|
||||
|
||||
For request-driven services, instrument **RED** on every endpoint and every external dependency: **R**ate (requests/sec), **E**rrors (failure rate), **D**uration (latency histogram, not average). For resources (queues, pools, hosts), use **USE**: **U**tilization, **S**aturation, **E**rrors.
|
||||
|
||||
As with tracing, the vendor-neutral path is the OpenTelemetry metrics API (same SDK and context as step 5). The example below uses Prometheus' `prom-client` — one common backend choice, not the only one; the RED/USE and cardinality rules are identical either way.
|
||||
|
||||
```typescript
|
||||
import { Histogram } from 'prom-client';
|
||||
|
||||
const httpDuration = new Histogram({
|
||||
name: 'http_request_duration_seconds',
|
||||
help: 'HTTP request duration',
|
||||
labelNames: ['method', 'route', 'status_class'], // '2xx', not '200'
|
||||
buckets: [0.05, 0.1, 0.25, 0.5, 1, 2.5, 5],
|
||||
});
|
||||
```
|
||||
|
||||
**Cardinality is the failure mode.** Every unique label combination is a separate time series. Labels must come from small, fixed sets (route template, status class, provider name). Never use user IDs, raw URLs, error messages, or other unbounded values as labels — that belongs in logs and traces.
|
||||
|
||||
```
|
||||
OK as label: route="/api/tasks/:id" status_class="5xx" provider="stripe"
|
||||
NEVER a label: user_id, email, request_id, full URL, error message text
|
||||
```
|
||||
|
||||
Track averages never, percentiles always: an average hides the 1% of users having a terrible time. Use histograms and read p50/p95/p99.
|
||||
|
||||
### 5. Distributed tracing
|
||||
|
||||
Use OpenTelemetry — it's the vendor-neutral standard, and auto-instrumentation covers HTTP, gRPC, and common DB clients with near-zero code:
|
||||
|
||||
```typescript
|
||||
// tracing.ts — must be imported before anything else
|
||||
import { NodeSDK } from '@opentelemetry/sdk-node';
|
||||
import { getNodeAutoInstrumentations } from '@opentelemetry/auto-instrumentations-node';
|
||||
|
||||
const sdk = new NodeSDK({
|
||||
serviceName: 'checkout-service',
|
||||
instrumentations: [getNodeAutoInstrumentations()],
|
||||
});
|
||||
sdk.start();
|
||||
```
|
||||
|
||||
Add manual spans only around meaningful internal units of work (e.g., `applyDiscounts`, `chargeProvider`) and attach the attributes on-call will filter by. Propagate context across every async boundary — HTTP headers, queue message metadata — or the trace dies at the gap. Sample head-based at a low rate by default; keep 100% of errors if your backend supports tail sampling.
|
||||
|
||||
### 6. Alerting
|
||||
|
||||
Alert on **symptoms users feel**, not on causes:
|
||||
|
||||
```
|
||||
SYMPTOM (page-worthy): CAUSE (dashboard, not a page):
|
||||
error rate > 1% for 5 min CPU at 85%
|
||||
p99 latency > 2s one pod restarted
|
||||
queue age > 10 min disk at 70%
|
||||
```
|
||||
|
||||
Cause-based alerts fire when nothing is wrong and miss failures you didn't predict. Symptom-based alerts fire exactly when users are hurt, regardless of the cause.
|
||||
|
||||
Rules for every alert you create:
|
||||
|
||||
1. **It must be actionable.** If the response is "ignore it, it self-heals", delete the alert.
|
||||
2. **It links to a runbook** — even three lines: what it means, first query to run, escalation path.
|
||||
3. **It has a threshold and duration** justified by the SLO or by historical data, not by a guess.
|
||||
4. Use two severities only: **page** (user-facing, act now) and **ticket** (degradation, act this week). A third tier becomes noise that trains people to ignore everything.
|
||||
|
||||
### 7. Verify the telemetry itself
|
||||
|
||||
Instrumentation is code; it can be wrong. Before calling the work done, trigger the paths and look at the actual output:
|
||||
|
||||
- Force an error in staging → find it in the logs by `requestId`, confirm fields are structured (not `[object Object]`)
|
||||
- Send test traffic → confirm metric series appear with the expected labels and sane values
|
||||
- Follow one request across services in the tracing UI → no broken spans
|
||||
- Fire each new alert once (lower the threshold temporarily) → confirm it reaches the right channel and the runbook link works
|
||||
|
||||
## Common Rationalizations
|
||||
|
||||
| Rationalization | Reality |
|
||||
|---|---|
|
||||
| "I'll add logging after it works" | "After" becomes "after the first incident", which is the most expensive moment to discover you're blind. Instrument as you build. |
|
||||
| "More logs = more observability" | Unstructured noise makes incidents slower, not faster. Three queryable events beat three hundred prose lines. |
|
||||
| "console.log is fine for now" | Unstructured output can't be filtered, correlated, or alerted on. The structured logger costs five extra minutes once. |
|
||||
| "We can just look at the dashboards when something breaks" | Dashboards built without defined questions show you everything except the answer. Start from on-call questions. |
|
||||
| "Alert on everything important, we'll tune later" | A noisy pager trains people to ignore it. The tuning never happens; the missed real page does. |
|
||||
| "User ID as a metric label makes debugging easier" | It also makes your metrics backend fall over. High-cardinality lookups belong in logs and traces. |
|
||||
| "Tracing is overkill for our two services" | Two services already means cross-service latency questions logs can't answer. Auto-instrumentation makes the cost trivial. |
|
||||
|
||||
## Red Flags
|
||||
|
||||
- A feature PR with retries, queues, or external calls and zero new telemetry
|
||||
- Log lines built by string interpolation instead of structured fields
|
||||
- No correlation/request ID — each log line is an orphan
|
||||
- Metrics labeled with user IDs, raw URLs, or error message text (cardinality bomb)
|
||||
- Latency tracked as an average with no percentiles
|
||||
- Alerts that fire daily and get acknowledged without action
|
||||
- Alerts on causes (CPU, memory) paging humans while user-facing error rate is unmonitored
|
||||
- Secrets, tokens, or full request bodies appearing in logs
|
||||
- "It works on my machine" as the only evidence a production feature is healthy
|
||||
|
||||
## Verification
|
||||
|
||||
After instrumenting a feature, confirm:
|
||||
|
||||
- [ ] The on-call questions for this feature are written down, and each signal maps to one
|
||||
- [ ] All log output is structured (JSON), with stable event names and a correlation ID on every line
|
||||
- [ ] No secrets, tokens, or unredacted PII in any log line (spot-check actual output)
|
||||
- [ ] RED metrics exist for every new endpoint and every external dependency, with bounded label sets
|
||||
- [ ] Latency is a histogram; p95/p99 are queryable
|
||||
- [ ] A single request can be followed end-to-end in the tracing UI without broken spans
|
||||
- [ ] Every new alert is symptom-based, has a runbook link, and was test-fired once
|
||||
- [ ] An induced failure in staging was located via telemetry alone, without reading the source
|
||||
|
||||
For the at-a-glance version of this list, including the pre-launch instrumentation gate, see `references/observability-checklist.md`.
|
||||
|
||||
## Limitations
|
||||
|
||||
- Use this skill only when the task clearly matches its upstream source and local project context.
|
||||
- Verify commands, generated code, dependencies, credentials, and external service behavior before applying changes.
|
||||
- Do not treat examples as a substitute for environment-specific tests, security review, or user approval for destructive or costly actions.
|
||||
Reference in New Issue
Block a user