.env.example

# =============================================================================
# Environment Variables for Test-2 CDC Pipeline
# =============================================================================
# This file contains all configuration for the CDC pipeline.
# DO NOT commit .env file to version control in production!
# 
# SETUP INSTRUCTIONS:
# 1. Copy this file: cp .env.example .env
# 2. Fill in actual values for your environment
# 3. Start services: docker compose up -d
# 
# MONITORING ACCESS:
# - Redpanda Console: http://localhost:8080 (no login)
# - Adminer: http://localhost:8090
#   PostgreSQL: postgres / postgres / test-2_db
# =============================================================================

# ===========================================================================
# Source Database Configuration (POSTGRES)
# ===========================================================================
POSTGRES_SOURCE_HOST=
POSTGRES_SOURCE_PORT=5432
POSTGRES_SOURCE_USER=
POSTGRES_SOURCE_PASSWORD=
POSTGRES_SOURCE_DB=

# ===========================================================================
# Target PostgreSQL Configuration
# ===========================================================================
POSTGRES_LOCAL_USER=postgres
POSTGRES_LOCAL_PASSWORD=postgres
POSTGRES_LOCAL_DB=test-2_db
POSTGRES_PORT=5432

# ===========================================================================
# Redpanda/Kafka Configuration
# ===========================================================================
REDPANDA_BROKERS=redpanda:9092
REDPANDA_SCHEMA_REGISTRY=http://redpanda:8081

# ===========================================================================
# CDC Pipeline Configuration
# ===========================================================================
CDC_BUFFER_SIZE=1000
CDC_BATCH_TIMEOUT=5s
CDC_MAX_IN_FLIGHT=64

# Sink runtime batching (used by Bento sql_insert in generated sink pipeline)
# Usage:
# - Set values in .env or deployment environment and restart sink runtime.
# - No pipeline regeneration is needed for these runtime settings.
# Best practices:
# - Keep SINK_BATCH_COUNT moderate first (100 default), then tune with metrics.
# - Prefer 2s-10s window for SINK_BATCH_PERIOD to balance latency and throughput.
SINK_BATCH_COUNT=100
SINK_BATCH_PERIOD=5s

# Source runtime polling/batching (used by generated source pipeline)
# Usage:
# - Adjust at runtime via environment variables; no regeneration needed.
# Best practices:
# - Keep MSSQL_CDC_SELECT_TOP conservative first (100000 default) and increase only if DB load allows.
# - Keep SOURCE_BATCH_PERIOD short (around 1s-5s) to avoid stale CDC latency.
MSSQL_CDC_SELECT_TOP=100000
SOURCE_BATCH_COUNT=100
SOURCE_BATCH_PERIOD=1s