Skip to content

PERFORMANCE_DASHBOARD_QUICKSTART.md

Latest commit

 

History

History
281 lines (207 loc) · 5.89 KB

PERFORMANCE_DASHBOARD_QUICKSTART.md

File metadata and controls

281 lines (207 loc) · 5.89 KB

Performance Dashboard - Quick Start Guide

Get the ThemisDB Performance Dashboard up and running in 5 minutes!

Prerequisites

  • Docker & Docker Compose installed
  • ThemisDB repository cloned
  • (Optional) ThemisDB built with benchmarks

Step 1: Start Dashboard Stack (2 minutes)

# Navigate to grafana directory
cd grafana

# Start Grafana + Prometheus
docker-compose up -d

# Verify services are running
docker-compose ps

Expected output:

NAME                COMMAND                  STATUS    PORTS
grafana             "/run.sh"                Up        0.0.0.0:3000->3000/tcp
prometheus          "/bin/prometheus --c…"   Up        0.0.0.0:9090->9090/tcp

Step 2: Access Dashboard (30 seconds)

  1. Open browser: http://localhost:3000
  2. Login: admin / admin
  3. (Optional) Change password or skip
  4. Navigate to: Dashboards → ThemisDB Performance Dashboard

That's it! 🎉 Dashboard is ready.

Step 3: Run Benchmarks (2 minutes)

# Option A: Quick test with example data
cd benchmarks
python3 performance_tracker.py \
  --results examples/sample_results.json \
  --storage performance_data

# Option B: Run actual benchmarks
cd build
./bench_crud --benchmark_format=json --benchmark_out=results.json
cd ..
python3 benchmarks/performance_tracker.py \
  --results build/results.json \
  --storage benchmarks/performance_data

Step 4: View Results (1 minute)

Refresh the Grafana dashboard. You should see:

  • ✅ Throughput metrics populated
  • ✅ Latency charts showing data
  • ✅ No critical regressions (first run)

Quick Commands Reference

Start/Stop Dashboard

# Start
cd grafana && docker-compose up -d

# Stop
docker-compose down

# Restart
docker-compose restart

# View logs
docker-compose logs -f grafana

Track Benchmarks

# Track results
python3 benchmarks/performance_tracker.py \
  --results <path-to-results> \
  --storage benchmarks/performance_data \
  --branch $(git branch --show-current) \
  --commit $(git rev-parse --short HEAD)

Manage Baselines

# Save baseline
python3 benchmarks/baseline_manager.py save \
  --results build/benchmark_results \
  --branch main \
  --version $(cat VERSION) \
  --commit $(git rev-parse HEAD)

# List baselines
python3 benchmarks/baseline_manager.py list

# Load baseline
python3 benchmarks/baseline_manager.py load --branch main

Detect Regressions

# Compare with baseline
python3 benchmarks/performance_regression_detector.py \
  --baseline benchmarks/baselines/main/latest.json \
  --current build/results.json \
  --output report.txt

Troubleshooting

Dashboard shows "No Data"

Check 1: Is Prometheus running?

curl http://localhost:9090/-/healthy
# Should return: Prometheus is Healthy.

Check 2: Are metrics being generated?

ls -la benchmarks/performance_data/metrics/
# Should see: benchmarks.prom

Check 3: Prometheus datasource configured?

  • Go to Grafana → Configuration → Data Sources
  • Click "Prometheus"
  • Click "Save & Test"
  • Should see: "Data source is working"

Fix: If still not working:

# Restart services
cd grafana
docker-compose restart

# Check Prometheus config
docker-compose exec prometheus cat /etc/prometheus/prometheus.yml

Can't login to Grafana

Default credentials:

  • Username: admin
  • Password: admin

If locked out:

# Reset admin password
docker-compose exec grafana grafana-cli admin reset-admin-password newpassword

# Or recreate containers
docker-compose down -v
docker-compose up -d

Benchmarks fail to build

Check build configuration:

# Ensure benchmarks are enabled
cmake -B build -DBUILD_BENCHMARKS=ON
cmake --build build --target bench_crud

Missing dependencies?

# Install Google Benchmark
sudo apt-get install libbenchmark-dev

Next Steps

1. Explore the Dashboard

  • Click through different panels
  • Try different time ranges (top-right)
  • Use template variables (Branch, Release, Hardware)
  • Export charts (Share → Export)

2. Set Up Alerts

# Configure Slack webhook
export SLACK_WEBHOOK_URL="https://hooks.slack.com/services/YOUR/WEBHOOK/URL"

# Edit prometheus.yml
vim grafana/prometheus.yml
# Add alerting section

# Restart Prometheus
cd grafana && docker-compose restart prometheus

3. CI/CD Integration

The performance checks run automatically on PRs. To test manually:

# Trigger workflow
gh workflow run performance-regression-check.yml

4. Create Baselines

# For main branch
git checkout main
# Run benchmarks...
python3 benchmarks/baseline_manager.py save \
  --results build/benchmark_results \
  --branch main \
  --version $(cat VERSION) \
  --commit $(git rev-parse HEAD)

Demo Mode

Want to see the dashboard with sample data?

# Generate sample metrics
python3 << 'EOF'
import random
import time

with open('benchmarks/performance_data/metrics/benchmarks.prom', 'w') as f:
    f.write('# HELP themisdb_benchmark_throughput_ops Throughput\n')
    f.write('# TYPE themisdb_benchmark_throughput_ops gauge\n')
    
    for i in range(10):
        throughput = random.randint(40000, 50000)
        f.write(f'themisdb_benchmark_throughput_ops{{benchmark="crud",branch="main"}} {throughput}\n')
        time.sleep(0.1)
EOF

# Prometheus will scrape these metrics automatically

Resources

Support

Having issues?

  1. Check Troubleshooting section above
  2. Search GitHub Issues
  3. Create new issue with performance label

Total Time: ~5 minutes ⏱️
Difficulty: Easy 🟢
You're ready to monitor performance! 🎯