DBChaos

Stress-test your database with pre-defined queries, generate synthetic data in your database. Validate slow and expensive queries that breaks your database.

Use GPT to generate synthetic data for you database.

Features

• Synthetic Event Generation
• Synthentic Data Generation

Installation

go install github.com/adaptive-scale/[email protected]

Supported Databases

Database	Synthetic Event Generation	Synthetic Data Generation
Postgres	✅	✅
MySQL	✅	✅
SQL Server	✅	✅
MongoDB	✅	⛔

Synthetic Event Generation

With DBChaos, you can run parallel queries on your target database. There two ways you could do that - test and scenarios.

With test, you can run single query on your target database. It would run the query parallely for the given amount of time. With Scenario, you could run multiple queries with different timeout and rates creating diverse load patterns.

We are planning to add more features around creating various load patterns.

Run your first test

Create a file named config.yaml with the following content:

dbType: postgres
connection: "host=localhost port=5432 user=postgres password=postgres dbname=postgres sslmode=disable"
query: |
  SELECT pg_database.datname as "Database", pg_size_pretty(pg_database_size(pg_database.datname)) as "Size"
  FROM pg_database;
parallelRuns: 100
runFor: 30m

For MongoDB, the connection string should be in the following format:

dbType: mongodb
connection: "mongodb://root:example@localhost:27017/"
query: |
    {"insert": "users", "documents": [{ "user": "abc123", "status": "A" }]}
parallelRuns: 100
runFor: 30m
dbName: users

To run the above config file:

dbchaos runTest 

Run bunch of queries in parallel

Create a file called scenario.yaml with the following content:

dbType: mysql
connection: "root:root@tcp(host:port)/db"
scenarios:
  - query: select * from information_schema.statistics
    parallelRuns: 10000
    runFor: 15m
  - query: |
      SELECT table_schema "Database", ROUND(SUM(data_length + index_length) / 1024 / 1024, 2) "Size (MB)"
      FROM information_schema.tables
      GROUP BY table_schema;
    parallelRuns: 10000
    runFor: 15m

To run the above scenario file:

dbchaos runScenario 

For MongoDB Specific, an example scenario.yaml file would look as follows :

dbType: mongodb
connection: "mongodb://root:example@localhost:27017/"
scenarios:
  - query: '{"insert": "users", "documents": [{ "user": "abc123", "status": "A" }]}'
    parallelRuns: 10000
    runFor: 15m
dbName: users   #(MongoDB only)

Synthetic Data Generation

DBChaos can generate full schema and synthetic data for your database. In DBChaos, there are two kinds of data generation techniques - Static and GPT-based.

In static data generation, dbchaos randomly generates schema, schema name, column names and data. It comes very handy and is inexpensive if you want to create huge schemas and generate large amount data. For instance, At Adaptive we use this is create unrealistic sized databases and schema to load testing our services and processes.

In GPT based data generation, you can create hyper-realistic databases and data. However, you would need an API key from OPENAI as well as it will cost you credits if you which to generate huge amount to data. We have tried to build a known schema cache in the product, which we will keep improving as well built out more features.

Static Data Generation

A configuration for static generation looks as follows:

connection: 
  dbType: postgres
  connection: "host=localhost port=5432 user=postgres password=postgres dbname=postgres sslmode=disable"
dryRun: false
schema: 
  numberOfSchema: 10
  generateTables: true
  language: en
tables:
  numberOfTables: 10
  minColumns: 5
  maxColumns: 10
  populateTable: true
rows:
  minRows: 100
  maxRows: 1000

Save above config as config.yaml and run the following command:

dbchaos generate

GPT-based Synthetic Data Generation

Configuration for GPT-based synthetic data looks as follows:

connection: 
  dbType: postgres
  connection: "host=localhost port=5432 user=postgres password=postgres dbname=postgres sslmode=disable"
dryRun: false
provider: openai
model: gpt-3.5-turbo
schema_type: webshop # can be anything word like ecommerce, webshop, hospital etc

You have to set your OpenAI API key as an environment variable OPENAI_API_KEY.

Save above config as config.yaml and run the following command:

dbchaos generateWithLLM

This will generate the schemas, insert commands and persist it in the database.