docs: Add CONFIG.md and UPGRADE_GUIDE.md with public-friendly URLs

- Add configuration guide for environment variables
- Add agent upgrade guide
- Use example.com instead of company-specific URLs
- No sensitive information included
This commit is contained in:
taylanbakircioglu
2025-11-12 14:15:51 +03:00
parent b3b0544b11
commit 73bed7811f
2 changed files with 448 additions and 0 deletions
+276
View File
@@ -0,0 +1,276 @@
# Configuration Guide - HAProxy OpenManager
## 📋 Yapılandırma Dosyaları
### Dosya Yapısı
```
haproxy-openmanager/
├── .env.template # ✅ Template (GIT'e commit edilir)
├── .env # ❌ Gerçek config (GIT'e commit EDİLMEZ)
├── .gitignore # .env dosyalarını korur
└── CONFIG.md # Bu dosya
```
## 🎯 Quick Start
### 1. Template'den Config Oluşturma
```bash
# Template'i kopyala
cp .env.template .env
# Gerçek değerleri düzenle
nano .env
```
### 2. Örnek Yapılandırmalar
#### Development (Local)
```bash
# .env dosyası
PUBLIC_URL=http://localhost:8000
MANAGEMENT_BASE_URL=http://localhost:8000
DEBUG=True
LOG_LEVEL=DEBUG
```
#### Staging
```bash
# .env dosyası veya K8s ConfigMap
PUBLIC_URL=https://haproxy-staging.company.com
MANAGEMENT_BASE_URL=https://haproxy-staging.company.com
DEBUG=False
LOG_LEVEL=INFO
```
#### Production (OpenShift)
```bash
# K8s ConfigMap: k8s/manifests/07-configmaps.yaml
data:
PUBLIC_URL: 'https://haproxy-manager.example.com'
MANAGEMENT_BASE_URL: 'https://haproxy-manager.example.com'
DEBUG: 'False'
LOG_LEVEL: 'INFO'
```
## 🔐 Güvenlik
### Hassas Bilgiler
`.env` dosyası hassas bilgiler içerir:
- ❌ Database şifreleri
- ❌ Secret key'ler
- ❌ API token'ları
**Bu yüzden**:
-`.env``.gitignore`'da (commit edilmez)
-`.env.template` → Git'e commit edilir (örnek değerler)
- ✅ Production'da: Kubernetes Secrets kullan
### .gitignore Kontrolü
```bash
# .env dosyalarının ignore edildiğini kontrol et
grep "^\.env" .gitignore
# Çıktı olmalı:
# .env
# .env.local
# .env.development.local
# .env.test.local
# .env.production.local
```
## 📚 Environment Variable Detayları
### PUBLIC_URL
**Ne İşe Yarar**: Agent kurulum script'lerinde kullanılır
**Örnekler**:
```bash
# Development
PUBLIC_URL=http://localhost:8000
# Production
PUBLIC_URL=https://haproxy-manager.company.com
# OpenShift
PUBLIC_URL=https://haproxy-manager.example.com
```
**Nasıl Kullanılır**:
1. Agent Management sayfasından "Generate Install Script"
2. Script içinde `{{MANAGEMENT_URL}}` bu değerle değiştirilir
3. Agent bu URL'ye bağlanarak backend'i dinler
### REACT_APP_API_URL
**Ne İşe Yarar**: Frontend'in backend'e bağlanacağı URL
**Özel Durum**:
```bash
# Boş bırakılırsa → Auto-detect (production için önerilen)
REACT_APP_API_URL=
# Development için explicit
REACT_APP_API_URL=http://localhost:8000
```
**Auto-detect Mantığı**:
```javascript
// frontend/src/utils/api.js
if (process.env.REACT_APP_API_URL) {
return process.env.REACT_APP_API_URL;
}
// Production'da same-origin kullan
if (window.location) {
return `${window.location.protocol}//${window.location.hostname}`;
}
```
## 🚀 Deployment Senaryoları
### Docker Compose
```bash
# 1. .env dosyası oluştur
cp .env.template .env
# 2. Değerleri düzenle
nano .env
# 3. Başlat
docker-compose up -d
# 4. Kontrol et
docker-compose logs backend | grep "PUBLIC_URL"
```
### Kubernetes/OpenShift
```bash
# 1. ConfigMap'i düzenle
vim k8s/manifests/07-configmaps.yaml
# 2. Apply
kubectl apply -f k8s/manifests/
# 3. Kontrol et
kubectl get configmap backend-config -n haproxy-manager -o yaml
```
### Manuel (Development)
```bash
# 1. Backend
cd backend
cp ../.env.template .env
export $(cat .env | xargs)
uvicorn main:app --reload
# 2. Frontend (başka terminal)
cd frontend
export REACT_APP_API_URL=http://localhost:8000
npm start
```
## 🔧 Troubleshooting
### Agent Script'inde Yanlış URL
**Sorun**: Agent script hala eski URL içeriyor
**Çözüm**:
```bash
# 1. Backend'deki değeri kontrol et
docker exec haproxy-openmanager-backend env | grep PUBLIC_URL
# 2. Container'ı yeniden başlat
docker-compose restart backend
# 3. Yeni script oluştur
# UI'dan tekrar "Generate Install Script"
```
### Frontend Backend'e Bağlanamıyor
**Sorun**: CORS hatası veya connection refused
**Çözüm**:
```bash
# 1. Frontend config'i kontrol et
docker exec haproxy-openmanager-frontend env | grep REACT_APP_API_URL
# 2. Browser console'da kontrol et
# [API Config] Base URL: http://localhost:8000
# 3. Network sekmesinde request URL'i kontrol et
```
## 📖 En İyi Pratikler
### ✅ YAPILMASI GEREKENLER
1. **Her ortam için ayrı değerler**
```
Dev: PUBLIC_URL=http://localhost:8000
Staging: PUBLIC_URL=https://staging.company.com
Prod: PUBLIC_URL=https://prod.company.com
```
2. **Template'i güncelle**
- Yeni variable eklendiğinde `.env.template`'e ekle
- Dokümantasyon ile birlikte
3. **Secrets kullan (Production)**
```yaml
# K8s Secret
apiVersion: v1
kind: Secret
metadata:
name: backend-secret
data:
SECRET_KEY: <base64-encoded>
```
### ❌ YAPILMAMASI GEREKENLER
1. **`.env` dosyasını commit etmeyin**
```bash
# Yanlış!
git add .env
# Doğru!
git add .env.template
```
2. **Production secret'larını template'e koymayın**
```bash
# .env.template içinde YANLIŞLAR:
SECRET_KEY=actual-production-secret-12345 ❌
DATABASE_URL=postgresql://admin:realpass@prod-db ❌
# Doğru:
SECRET_KEY=your-secret-key-change-this-in-production ✅
DATABASE_URL=postgresql://user:pass@host:5432/db ✅
```
3. **Hardcoded URL kullanmayın**
```python
# Yanlış!
MANAGEMENT_URL = "https://my-server.com" ❌
# Doğru!
MANAGEMENT_URL = os.getenv("PUBLIC_URL") ✅
```
## 🆘 Yardım
Sorularınız için:
- 📘 Bu dosya: `CONFIG.md`
- 📗 Ana dokümantasyon: `README.md`
- 📙 Template: `.env.template`
+172
View File
@@ -0,0 +1,172 @@
# Agent Upgrade Guide - Dashboard Stats Fix
## Problem
Dashboard showing 0 metrics after agent auto-upgrade due to missing environment variables (`SOCAT_BIN`, `STATS_SOCKET_PATH`) when agent restarts in daemon mode.
## Solution
Implemented **lazy initialization** in both `get_haproxy_stats_csv()` and `get_server_statuses()` functions. These functions now initialize their dependencies on first call, making them completely independent of global variable initialization.
## Deployment Steps
### Step 1: Wait for Pipeline ⏳
```bash
# Pipeline is currently running after git push
# Check status: https://[your-azure-devops]/pipeline
# Wait for deployment to complete (~3-5 minutes)
```
### Step 2: Verify Backend Deployment ✅
```bash
# Check backend logs for successful deployment
kubectl logs -f deployment/backend -n haproxy-manager | head -20
# Expected: New pod started with latest code
```
### Step 3: Update Agent Version in UI 🔄
**Option A: Automatic Script Sync (Recommended)**
```bash
# Get admin token
TOKEN=$(curl -k -s -X POST "https://haproxy-manager.example.com/api/auth/login" \
-H "Content-Type: application/json" \
-d '{"username": "admin", "password": "admin123"}' | jq -r '.access_token')
# Sync scripts from files to database (creates version 1.0.0)
curl -k -X POST "https://haproxy-manager.example.com/api/agents/sync-scripts-from-files" \
-H "Authorization: Bearer $TOKEN" | jq .
# Expected: {"status": "success", "synced": ["linux", "macos"]}
```
**Option B: Manual UI Update**
1. Go to **Agent Management** page
2. Click **Settings****Agent Versions**
3. For **Linux** platform:
- Click **Edit Script**
- Version: Keep current or increment (e.g., 1.0.3)
- Changelog: Add "Fixed stats collection after upgrade"
- Click **Save** (this syncs file content to database)
4. Repeat for **macOS** platform
### Step 4: Upgrade Agents 🚀
**Option A: UI (Single Agent)**
1. Go to **Agent Management** page
2. Select agent (e.g., `demo-agent`)
3. Click **Upgrade** button
4. Wait 30 seconds for agent to restart
**Option B: API (Batch Upgrade)**
```bash
# Get all agents
curl -k -s -X GET "https://haproxy-manager.example.com/api/agents" \
-H "Authorization: Bearer $TOKEN" | jq '.agents[] | {id, name, version}'
# Upgrade specific agent (replace {agent_id})
curl -k -X POST "https://haproxy-manager.example.com/api/agents/{agent_id}/upgrade" \
-H "Authorization: Bearer $TOKEN" | jq .
```
### Step 5: Verify Stats Collection 📊
**Backend Logs** (30 seconds after upgrade):
```bash
kubectl logs -f deployment/backend -n haproxy-manager | grep -E "haproxy_stats_csv|demo-agent"
# Expected logs:
# ✅ "Has haproxy_stats_csv: True"
# ✅ "CSV preview: IyBweG..." (base64 data)
# ✅ "STATS: Parsed 15 rows for cluster demo-cluster1"
```
**Agent Logs** (on agent server):
```bash
sudo tail -f /var/log/haproxy-agent/agent.log | grep STATS
# Expected logs:
# ✅ "STATS: Initialized socat: /usr/bin/socat"
# ✅ "STATS: Using default socket path: /var/run/haproxy/admin.sock"
# ✅ "STATS: Fetched CSV: 2121 bytes, 9 lines, base64: 2828 chars"
```
**Dashboard UI**:
1. Open **Dashboard** page
2. Refresh page (F5)
3. Check metrics:
- ✅ Frontend/Backend filters populated
- ✅ Overview metrics showing real data (not 0)
- ✅ Charts showing data points
- ✅ "Waiting for Agent Data" warning gone
## Troubleshooting
### Issue: Backend still shows "Has haproxy_stats_csv: False"
**Cause**: Agent hasn't upgraded yet or using old script version
**Solution**:
```bash
# Check agent version on agent server
grep "AGENT_VERSION" /usr/local/bin/haproxy-agent | head -1
# Force agent restart
sudo systemctl restart haproxy-agent
# Check logs immediately
sudo tail -20 /var/log/haproxy-agent/agent.log
```
### Issue: "STATS: socat not available"
**Cause**: socat not installed
**Solution**:
```bash
# Install socat
sudo yum install -y socat # RHEL/CentOS
sudo apt install -y socat # Debian/Ubuntu
# Restart agent
sudo systemctl restart haproxy-agent
```
### Issue: "STATS: Socket not found: /var/run/haproxy/admin.sock"
**Cause**: HAProxy stats socket not configured
**Solution**:
```bash
# Check HAProxy config for stats socket
grep "stats socket" /etc/haproxy/haproxy.cfg
# Add if missing (in global section):
# stats socket /var/run/haproxy/admin.sock mode 666 level admin
# Reload HAProxy
sudo systemctl reload haproxy
```
## Verification Checklist
- [ ] Pipeline completed successfully
- [ ] Backend pod restarted with new code
- [ ] Agent scripts synced to database (version visible in UI)
- [ ] Agents upgraded to new version
- [ ] Backend logs show "Has haproxy_stats_csv: True"
- [ ] Agent logs show STATS messages
- [ ] Dashboard showing real metrics
- [ ] Charts populated with data
- [ ] No "Waiting for Agent Data" warning
## Next Upgrades
This fix is **permanent**. Future agent upgrades will NOT break stats collection because:
1. ✅ Functions are self-contained with lazy initialization
2. ✅ No dependency on global variable initialization order
3. ✅ Works in any restart scenario (systemd, daemon mode, upgrade)
4. ✅ Backward compatible with existing agents
**No manual intervention needed for future upgrades!** 🎉