Run Modes Overview

Sercha Core supports three run modes, enabling flexible deployment from single-container development to distributed production clusters.

Mode Comparison

Mode	HTTP Server	Task Queue	Scheduler	Use Case
all	Yes	Yes	Yes	Development, simple deployments
api	Yes	No	No	Horizontally scaled API tier
worker	No	Yes	Yes	Dedicated background processing

Configuration

Set the run mode via the RUN_MODE environment variable:

services:
  sercha-api:
    image: sercha-core:latest
    environment:
      RUN_MODE: api
      # ... other config

Value	Description
all	Combined mode (default)
api	HTTP server only
worker	Background processing only

Architecture

Shared Infrastructure

All containers connect to the same backing services:

Service	Purpose	Required
PostgreSQL	User data, documents, metadata	Yes
Vespa	Search index	Yes
Redis	Sessions, queue, distributed locks	No (falls back to PostgreSQL)

Graceful Shutdown

All modes respond to container stop signals:

1. Container runtime sends SIGTERM
2. In-flight requests complete (API) or current task finishes (Worker)
3. Connections drain within 30 seconds
4. Container exits cleanly

Next Steps

• Configuration - Environment variables reference
• API Mode - HTTP server configuration
• Worker Mode - Task processing and scheduling
• Scaling - Multi-instance deployment patterns