README
¶
DLG - Da Load Generator
dlg is a universal load generator that can be used to generate load for various protocols and services.
Supported Protocols
DLG supports the following protocols:
- HTTP/HTTPS - HTTP/HTTPS load testing with customizable headers, methods, and payloads
- gRPC - gRPC service load testing with metadata and connection options
- WebSocket - WebSocket connection testing with read/write operations
- GraphQL - GraphQL query execution with variables and custom headers
- MongoDB - MongoDB database operations (insert, find, update, delete, count, aggregate)
- Cassandra - Cassandra CQL queries with configurable consistency levels
- Elasticsearch - Elasticsearch index, search, and document operations
- InfluxDB - InfluxDB time-series data writes and queries
- Redis - Redis commands and operations
- SQL - MySQL, PostgreSQL, and ClickHouse database testing
- MQTT - MQTT publish/subscribe messaging with QoS levels
- ETCD - ETCD key-value operations
- DNS - DNS query load testing
- LDAP - LDAP authentication and query testing
- Memcache - Memcache get/set/delete operations
- SNMP - SNMP protocol testing
- SSH - SSH command execution
- UDP - UDP protocol testing
- DHCP4 - DHCP4 protocol testing
Features
- Multi-protocol support - Test multiple protocols in a single plan
- Staged execution - Organize tests into stages with configurable repeats and durations
- Concurrent execution - Run stages concurrently for maximum load
- Prometheus metrics - Built-in Prometheus metrics for monitoring
- Rate limiting - Control load with configurable rate limiters
- Flexible configuration - YAML-based configuration with extensive options
- MCP Server - Model Context Protocol server for AI agent integration
MCP Server for AI Agents
DLG includes a built-in Model Context Protocol (MCP) server that allows AI agents to use DLG for load testing. This enables AI-powered workflows to automatically generate and execute load tests.
Starting the MCP Server
dlg mcp
The MCP server runs over stdio and exposes the following tools:
- http_load_test - Generate HTTP load against a target URL
- redis_load_test - Generate Redis load against a server
- mongodb_load_test - Generate MongoDB load against a database
- postgres_load_test - Generate PostgreSQL load against a database
- websocket_load_test - Generate WebSocket load against an endpoint
- grpc_load_test - Generate gRPC load against a service
- run_load_plan - Execute a YAML-based load test plan
Example AI Agent Usage
AI agents can use these tools to perform load testing as part of automated workflows. For example:
- Performance testing during CI/CD pipelines
- Automated capacity planning
- Dynamic stress testing based on monitoring data
- Interactive load testing through natural language
MCP Integration
To integrate DLG with an AI agent:
- Start the MCP server:
dlg mcp - Configure your AI agent to connect to the MCP server via stdio
- The agent can now call load testing tools with natural language
Each tool returns Prometheus metrics from the load test execution.
Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func NewManagerRouter ¶
NewManagerRouter returns a new manager router.
Types ¶
type Manager ¶
type Manager interface {
executor.Plan
// Get is used to return a plan by name.
Get(context.Context, string) (*config.Plan, error)
// Add is used to add a new plan.
Add(context.Context, *config.Plan) error
// Delete is used to remove a plan.
Delete(context.Context, string) error
// Plans returns a list of all known plans.
Plans(context.Context) ([]*config.Plan, error)
}
Manager is used for scheduling execution plans.
Source Files
Directories
Click to show internal directories.
Click to hide internal directories.