# Performance monitoring

What to watch on your node, and how to tune it. Uptime and task completion feed your [TrustFingerprint](../protocol/trustfingerprint/README.md) score and your share of node rewards.

> Rewards are not being paid yet. See [Testnet Status](../introduction/testnet-status.md).

## Recommended tools

- **Prometheus:** metrics collection and storage
- **Grafana:** dashboards
- **Alertmanager:** alert routing
- **Node Exporter:** system metrics

## Targets

| Metric | Target | Critical Threshold |
|--------|--------|-------------------|
| **CPU Usage** | <70% average | >90% sustained |
| **Memory Usage** | <80% | >95% |
| **Disk I/O** | <80% capacity | >95% capacity |
| **Network Bandwidth** | <70% capacity | >90% capacity |

| Metric | Target | Impact |
|--------|--------|--------|
| **Uptime** | 99%+ | Direct reward multiplier |
| **Response Time** | <100ms | TrustFingerprint score |
| **Task Completion Rate** | 99%+ | Reward eligibility |
| **Peer Connections** | 20+ | Network health |

## What the score measures

1. **Uptime** (40% weight): historical availability
2. **Performance** (30% weight): task completion speed and accuracy
3. **Reliability** (20% weight): consistency over time
4. **Participation** (10% weight): governance and network engagement

## Dashboards

- **System:** CPU usage by core, memory and swap, disk I/O and space, network traffic, system load average
- **Node:** uptime, task completion rate, reward earnings, TrustFingerprint score, peer connections, block and transaction processing

## Alerts

Set alerts for high CPU or memory usage, low disk space, network connectivity issues, node offline, low TrustFingerprint score and missed tasks.

- **Critical, act now:** node offline >5 minutes, CPU >95% for >10 minutes, memory >98%, disk space <5%, network disconnected
- **Warning, investigate:** CPU >80% for >1 hour, memory >85%, disk space <20%, TrustFingerprint score declining, task completion rate <95%
- **Info, awareness only:** software updates available, reward distribution completed, governance proposals active, network announcements

## Tuning

```bash
# Check CPU frequency scaling
cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor

# Set to performance mode
echo performance | sudo tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
```

```bash
# Adjust swappiness
sudo sysctl vm.swappiness=10

# Clear cache if needed
sudo sync && sudo sysctl -w vm.drop_caches=3
```

```bash
# Increase network buffer sizes
sudo sysctl -w net.core.rmem_max=134217728
sudo sysctl -w net.core.wmem_max=134217728
```

```bash
# Enable TRIM for SSDs
sudo fstrim -v /

# Check disk health
sudo smartctl -a /dev/sda
```

## Fix common problems

- **High CPU usage:** find the process with `top` or `htop`. Check for runaway processes and verify the node configuration. Consider a hardware upgrade.
- **Memory leak:** watch memory over time. Restart the node if it grows continuously. Check for software updates and report the issue if it persists.
- **Network latency:** test with `ping 8.8.8.8` and `speedtest-cli`. Verify the router configuration. Consider an ISP upgrade.
- **Low TrustFingerprint score:** review historical performance, find the periods of downtime, check your task completion rate and improve system reliability.

## Routine

- **Daily:** check the dashboard for alerts, verify the node is online, review overnight performance, check reward earnings.
- **Weekly:** review performance trends, update software if needed, check disk space, back up configuration.
- **Monthly:** analyze TrustFingerprint trends, tune the system, review alerts, plan hardware upgrades, rerun the benchmarks below.

## Benchmarks

Record a baseline after setup. Run it again monthly to detect degradation.

```bash
# CPU benchmark
sysbench cpu run

# Memory benchmark
sysbench memory run

# Disk benchmark
fio --name=random-write --ioengine=libaio --rw=randwrite --bs=4k --size=1G

# Network benchmark
iperf3 -c speedtest.net
```

## Report an issue

Collect diagnostics:

```bash
# System info
uname -a
lscpu
free -h
df -h

# Node logs
journalctl -u noderr-node -n 1000 > node-logs.txt

# Metrics export
curl http://localhost:8080/metrics > metrics.txt
```

Then report in [Discord](https://discord.com/invite/bPznF37J2e) or on [GitHub](https://github.com/Noderrxyz).

## Next

- [Setup Guide](./setup-guide.md)
- [Troubleshooting](../resources/troubleshooting.md)
- [Compare the tiers](./README.md)