Infrastructure Components

FeatBit Application Services relies on three critical infrastructure components: Database, Message Queue, and Caching.

Database

The main database, used to store all feature flags, segments, end-users, etc. It corresponds to the DbProvider environment variable for API, ELS, and DAS.

Available options include:

• Primary Database: MongoDB or Postgres for storing feature flags, segments, end-users data, etc.
• Analytics Database: Standalone and Standard version will use the primary database, while Professional version leverages ClickHouse for high-performance analytics on large datasets

Message Queue

The message queue system ensures reliable communication between FeatBit services and enables real-time updates to SDKs. It corresponds to the MqProvider environment variable for API and ELS.

There are two types of messages across the system:

• Real-time Messages: Feature flag and segment changes that require immediate delivery
• Job Messages: Background tasks processed sequentially based on timestamp

Available options include:

• Postgres: Uses LISTEN/NOTIFY for real-time messages and batch polling with timestamp ordering for job processing
• Redis: Uses PUB/SUB for real-time messages and list operations (RPUSH/LPOP) for job processing
• Kafka: Uses topics and partitions for both real-time messages and job processing

Caching

The caching layer is used to boost the performance of the ELS. It corresponds to the CacheProvider environment variable for API and ELS.

Available options include:

• Redis: ELS reads data from Redis and API ensures data in Redis is up to date.
• None: ELS reads data directly from the database.

How to Choose

When selecting infrastructure providers, consider:

• Existing Infrastructure: Leverage your organization's existing technology investments
• Operational Familiarity: Choose technologies your team is comfortable maintaining
• Performance Expectations: High traffic loads benefit from Kafka messaging and ClickHouse analytics