Metrics

The RocksDB Server includes a metrics server that exposes various performance and operational metrics for monitoring and analysis. The metrics are exposed in the Prometheus format.

Enabling Metrics Server

To enable the metrics server, use the --metrics parameter when starting the RocksDB Server. This parameter enables the metrics server, which will run on the same address as the RocksDB Server.

Example:

rocksdb_server --dbpath ./db_test --address 127.0.0.1:12345 --log-level info --metrics true

Available Metrics

The following metrics are exposed by the server:

• requests (Counter): The total number of requests received by the server.
• request_success_total (Counter): Total number of successful requests.
• request_failure_total (Counter): Total number of failed requests.
• request_duration_seconds (Histogram): The duration of the requests in seconds.
• response_speed_bytes (Counter): The speed of the response in bytes.
• cache_hits_total (Counter): The total number of cache hits.
• cache_misses_total (Counter): The total number of cache misses.
• cache_set_total (Counter): The total number of cache sets.
• active_connections (Gauge): The number of active connections.
• memory_usage_bytes (Gauge): Current memory usage of the process in bytes.
• cpu_usage_percentage (Gauge): Current CPU usage of the process in percentage.
• process_uptime_seconds (Gauge): Uptime of the process in seconds.

Accessing Metrics

Once the metrics server is enabled, you can access the metrics by navigating to the /metrics endpoint of your server in your web browser or using a tool like curl.

Example:

curl http://127.0.0.1:12345/metrics

Prometheus Integration

To integrate the RocksDB Server metrics with Prometheus, add the following job to your Prometheus configuration file:

scrape_configs:
  - job_name: 'rocksdb_server'
    static_configs:
      - targets: ['127.0.0.1:12345']

Replace 127.0.0.1:12345 with the actual address and port where your RocksDB Server is running.

Example Prometheus Configuration

Here is an example of a Prometheus configuration file that includes the RocksDB Server metrics:

global:
  scrape_interval: 15s
  evaluation_interval: 15s

scrape_configs:
  - job_name: 'rocksdb_server'
    metrics_path: '/metrics'
    static_configs:
      - targets: ['127.0.0.1:12345']

Save this configuration to a file, for example, prometheus.yml, and start Prometheus with the following command:

prometheus --config.file=prometheus.yml

This configuration will tell Prometheus to scrape the metrics from the RocksDB Server every 15 seconds.

Grafana Integration

To integrate the RocksDB Server metrics with Grafana, follow these steps:

1. Add Prometheus Data Source:

• Open Grafana in your web browser.
• Navigate to Configuration -> Data Sources.
• Click Add data source and select Prometheus.
• Set the URL to http://<PROMETHEUS_HOST>:<PROMETHEUS_PORT> (e.g., http://localhost:9090).
• Click Save & Test to verify the connection.

2. Create a Dashboard:

• Navigate to Create -> Dashboard.
• Click Add new panel.
• In the Query section, select Prometheus as the data source.
• Use the following queries to display the various metrics:

Example Grafana Dashboard Panels

{
  "panels": [
    {
      "title": "Total Requests",
      "type": "stat",
      "targets": [
        {
          "expr": "requests",
          "format": "time_series"
        }
      ]
    },
    {
      "title": "Successful Requests",
      "type": "stat",
      "targets": [
        {
          "expr": "request_success_total",
          "format": "time_series"
        }
      ]
    },
    {
      "title": "Failed Requests",
      "type": "stat",
      "targets": [
        {
          "expr": "request_failure_total",
          "format": "time_series"
        }
      ]
    },
    {
      "title": "Request Duration",
      "type": "graph",
      "targets": [
        {
          "expr": "histogram_quantile(0.95, sum(rate(request_duration_seconds_bucket[5m])) by (le))",
          "legendFormat": "95th percentile"
        },
        {
          "expr": "histogram_quantile(0.50, sum(rate(request_duration_seconds_bucket[5m])) by (le))",
          "legendFormat": "50th percentile"
        }
      ]
    },
    {
      "title": "Response Speed",
      "type": "graph",
      "targets": [
        {
          "expr": "rate(response_speed_bytes[5m])",
          "legendFormat": "Response Speed"
        }
      ]
    },
    {
      "title": "Cache Hits",
      "type": "stat",
      "targets": [
        {
          "expr": "cache_hits_total",
          "format": "time_series"
        }
      ]
    },
    {
      "title": "Cache Misses",
      "type": "stat",
      "targets": [
        {
          "expr": "cache_misses_total",
          "format": "time_series"
        }
      ]
    },
    {
      "title": "Cache Sets",
      "type": "stat",
      "targets": [
        {
          "expr": "cache_set_total",
          "format": "time_series"
        }
      ]
    },
    {
      "title": "Active Connections",
      "type": "stat",
      "targets": [
        {
          "expr": "active_connections",
          "format": "time_series"
        }
      ]
    },
    {
      "title": "Memory Usage",
      "type": "graph",
      "targets": [
        {
          "expr": "memory_usage_bytes",
          "legendFormat": "Memory Usage"
        }
      ]
    },
    {
      "title": "CPU Usage",
      "type": "graph",
      "targets": [
        {
          "expr": "cpu_usage_percentage",
          "legendFormat": "CPU Usage"
        }
      ]
    },
    {
      "title": "Process Uptime",
      "type": "stat",
      "targets": [
        {
          "expr": "process_uptime_seconds",
          "format": "time_series"
        }
      ]
    }
  ]
}

This JSON configuration defines a Grafana dashboard with panels for each metric exposed by the RocksDB Server. To use it:

1. Copy the JSON configuration.
2. In Grafana, navigate to Create -> Import.
3. Paste the JSON into the Import via panel json field.
4. Click Load and configure the dashboard as needed.