Synthetic Data Generator

A flexible synthetic data generator in Go, designed to produce realistic logs and metrics for testing and development.

Quick Start

Requires Go version 1.24+

• Clone the repository (alternatively download the latest release)
• Copy config.sample.yaml and edit as needed.
• Start the data generator with config file as the only parameter go run cmd/main.go --config ./config.yaml

Tip

Use --debug flag for debug level logs.

Configurations

Given below are the supported configuration options. Check config.sample.yaml for reference.

Input configurations

Given below are supported input types and their related environment variable overrides,

YAML Property	Environment Variable	Description
type	ENV_INPUT_TYPE	Specifies the input data type (e.g., LOGS, METRICS, ALB, NLB, VPC, CLOUDTRAIL, WAF).
delay	ENV_INPUT_DELAY	Delay between a data point. Accepts value in format like 5s (5 seconds), 10ms (10 milliseconds).
batching	ENV_INPUT_BATCHING	Set time delay between data batches. Accepts a time value similar to delay. Default is set to 0 (no batching). Note: Batching is most effective with bulk ingest endpoints like S3 and Firehose. For CloudWatch Logs, batching may not be suitable as it concatenates multiple log entries into single messages.
max_batch_size	ENV_INPUT_MAX_BATCH_SIZE	Set maximum byte size of a batch. Default is to ignore (no max size).
max_data_points	ENV_INPUT_MAX_DATA_POINTS	Set maximum amount of data points to generate. Default is to ignore (no max limit).
max_runtime	ENV_INPUT_MAX_RUNTIME	Set maximum duration load generator will run. Default is to ignore (no max limit).

Note about input type,

• LOGS : ECS (Elastic Common Schema) formatted logs based on zap
• METRICS: Generate metrics similar to a CloudWatch metrics entry
• ALB : Generate AWS ALB formatted logs with some random content
• NLB : Generate AWS NLB formatted logs with some random content
• VPC: Generate AWS VPC formatted logs with randomized content
• CLOUDTRAIL: Generate AWS CloudTrail formatted logs with randomized content. Data is generated for AWS S3 Data Event.
• WAF: Generate AWS WAF formatted logs with randomized content

Example:

input:
  type: LOGS             # Input type LOGS
  delay: 500ms           # 500 milliseconds between each data point
  batching: 10s          # Emit generated data batched within 10 seconds
  max_batch_size: 10000  # Limit maximum batch size to 10,000 bytes. The output is capped at 1000 bytes/second max
  max_data_points: 10000 # Exit input after generating 10,000 data points

Tip

When max_batch_size is reached, elapsed time for batching will be considered before generating new data

Output configurations

Given below are supported output types (environment variable ENV_OUT_TYPE),

• FILE: Output to a file
• FIREHOSE: Output to a Firehose stream
• CLOUDWATCH_LOG: Output to a CloudWatch log group
• S3: Output to a S3 bucket

Sections below provide output specific configurations

FILE

YAML Property	Environment Variable	Description
location	ENV_OUT_LOCATION	Output file location. Default to ./out. When batching, file suffix will increment with numbers (e.g., out_0, out_2).

Example:

output:
  type: FILE
  config:
    location: "./data"

S3

YAML Property	Environment Variable	Description
s3_bucket	ENV_OUT_S3_BUCKET	S3 bucket name (required).
compression	ENV_OUT_COMPRESSION	To compress or not the output. Currently supports gzip.
path_prefix	ENV_OUT_PATH_PREFIX	Optional prefix for the bucket entry. Default to logFile-.

Example:

output:
  type: S3
  config:
    s3_bucket: "testing-bucket"
    compression: gzip
    path_prefix: "datagen"

FIREHOSE

YAML Property	Environment Variable	Description
stream_name	ENV_OUT_STREAM_NAME	Firehose stream name (required).

Example:

output:
  type: FIREHOSE
  config:
    stream_name: "my-firehose-stream"

CLOUDWATCH_LOG

YAML Property	Environment Variable	Description
log_group	ENV_OUT_LOG_GROUP	CloudWatch log group name.
log_stream	ENV_OUT_LOG_STREAM	Log group stream name.

Example:

output:
  type: CLOUDWATCH_LOG
  config:
    logGroup: "MyGroup"
    logStream: "data"

Note

CloudWatch Logs API (PutLogEvents) is optimized for single log messages per API call. When batching is enabled, multiple log entries are concatenated into a single message, which may not be ideal for log analysis and searching in CloudWatch. For CloudWatch destinations, consider setting batching: 0s (no batching) or using a small delay without batching. Batching is more suitable for bulk ingest endpoints like S3 and Firehose.

CSP configurations

Currently, this project only support AWS Cloud Service Provider (CSP). Given below are available configurations,

YAML Property	Environment Variable	Description
region	AWS_REGION	Region to use by exporters. Default is us-east-1.
profile	AWS_PROFILE	Credential profile to use by exporters. Default is default.

Example:

aws:
  region: "us-east-1"
  profile: "default"

Example Configurations

1. Continuous Log Generation to a S3 bucket

Generate ECS-formatted logs every 2s, batch them in 10 seconds and forward to S3 bucket

input:
  type: LOGS
  delay: 2s
  batching: 10s
output:
  type: s3
  config:
    s3_bucket: "testing-bucket"

2. Continuous Log Generation with batching to a S3 bucket

Generate ALB logs. No delay between data points (continuous data generating). Limit batching to 10 seconds and max batch size is set to 10MB. This translates to ~1 MB/second data load. S3 files will be in gzip format.

input:
  type: ALB
  delay: 0s
  batching: 10s
  max_batch_size: 10000000
output:
  type: s3
  config:
    s3_bucket: "testing-bucket"
    compression: "gzip"

3. Limit data points

Generate VPC logs and limit to 2 data points. Then upload it to S3 in gzip format.

input:
  type: VPC
  delay: 1s
  max_data_points: 2
output:
  type: s3
  config:
    s3_bucket: "testing-bucket"
    compression: "gzip"

4. Limit runtime

Generate CLOUDTRAIL logs and limit to generator runtime of 5 minutes.

input:
  type: CLOUDTRAIL
  delay: 10us       # 10 microseconds between data points
  batching: 10s
  max_runtime: 5m   # 5 minutes
output:
  type: s3
  config:
    s3_bucket: "testing-bucket"
    compression: "gzip"