Skip to main content

Overview

Capturing metrics is not enough—you need a place to visualize trends and patterns over time. DynamoAI ships with Grafana configured by default, connected to a Prometheus instance that stores all your metrics as a data source.

About These Dashboards​

The dashboards described in this section are for System Metrics—specifically, API request metrics that capture HTTP-level statistics including request rates, response times, error rates, and bandwidth usage. These metrics provide insights into API performance and system health.

Note: For Application Metrics (such as DynamoGuard latency metrics), see Application Metrics.

How These Dashboards Work​

These Grafana dashboards query Prometheus for System Metrics data. The monitoring setup works as follows:

  1. API Layer: The DynamoAI API sends System Metrics (API request statistics) to the OpenTelemetry collector via OTLP (OpenTelemetry Protocol)
  2. OpenTelemetry Collector: The opentelemetry-collector receives and processes System Metrics from the API
  3. Prometheus: Prometheus is configured to scrape metrics from the OpenTelemetry collector for System Metrics
  4. Grafana Dashboards: These dashboards query Prometheus to visualize the metrics

In summary, the API only sends metrics to OpenTelemetry—it does not send metrics directly to Prometheus. Prometheus scrapes the metrics from the OpenTelemetry collector, and Grafana queries Prometheus to display the data in these dashboards.

Where can you view the dashboards?​

Endpoint​

The endpoint varies depending on your deployment configuration:

  • If ingress enabled:
    • <ui-endpoint>/grafana
    • Example: https://staging.dynamofl.com/grafana
  • If ingress disabled:
    • Domain-specific endpoint in deployment configuration
    • Example: https://grafana.dynamofl.com

Credentials​

Configured via the deployment config:

  • Default username: admin
  • Password: Contact your deployment team

Available Dashboards​

DynamoAI provides three pre-configured Grafana dashboards for comprehensive System Metrics (API request metrics) monitoring:

API Requests​

Purpose: System-wide overview of API health and performance across all endpoints.

Use when:

  • Daily operations monitoring and health checks
  • Incident response and troubleshooting
  • Quick assessment of overall API health

Key Metrics: Request rates, response times, error rates, endpoint popularity

API Path​

Purpose: Endpoint-specific analysis for individual API paths.

Use when:

  • Investigating specific endpoint issues
  • Monitoring latency-critical endpoints (e.g., DynamoGuard Analyze)
  • Comparing performance across different paths

Key Features: Path dropdown selector, filtered metrics per endpoint, detailed error breakdown

API Bandwidth​

Purpose: Data transfer monitoring and bandwidth consumption analysis.

Use when:

  • Tracking bandwidth usage and costs
  • Identifying endpoints with large payloads

Key Metrics: Bandwidth usage, payload sizes, throughput by endpoint and method

Dashboard Workflow​

Recommended workflow for monitoring:

  1. Start with API Requests - Get a high-level view of overall API health
  2. Drill down with API Path - Select specific problematic endpoints for detailed analysis
  3. Analyze bandwidth with API Bandwidth - Identify data-heavy endpoints

All dashboards update in real-time and support flexible time range selection for historical analysis.

Dashboard Configuration Files​

Download the Grafana dashboard JSON files for System Metrics (API request metrics) to import into your own Grafana instance:

Important Notes:

  • These dashboards require DynamoAI version 3.24 or later (metrics are only available from version 3.24)
  • Before importing, you'll need to replace the datasource UID placeholder in each JSON file with your Prometheus datasource UID
  • See the Dashboard Setup Guide for detailed instructions on configuring and importing these dashboards