Skip to main content

​
Overview

Automated failover handles message delivery by automatically switching to backup providers or channels when primary delivery methods fail. When a provider experiences outages, rate limiting, or timeouts, Courier automatically routes messages through alternative paths without manual intervention.

​
Key Concepts

​
Failover Triggers

Courier initiates failover when downstream providers return specific error conditions:
  • 408 Request Timeout: Provider response takes too long
  • 429 Too Many Requests: Rate limiting or throttling detected
  • >=500 Server Errors: Internal server errors, service unavailable

​
Failover Types

  • Provider Failover: Switch between different providers within the same channel (e.g., SendGrid → AWS SES for email)
  • Channel Failover: Switch between different communication channels (e.g., email → SMS → push)
  • Timeout-Based Failover: Automatically trigger failover based on response time thresholds

​
Configuration

​
Provider Failover

Set up multiple providers within a single channel to create redundancy at the provider level:

Provider Failover Example

Configure SendGrid as backup for AWS SES. If SES experiences an outage or rate limiting, Courier automatically switches to SendGrid for email delivery.
Email channel configuration showing AWS SES as primary provider and SendGrid as backup provider

Provider Failover Configuration

Configuration Options:
  • Template channel settings: Configure provider priority in the template designer
  • Send API: Use message.channels.[channel_name].routing_method property
Common Use Cases:
  • Email redundancy: SendGrid + AWS SES + Mailgun
  • SMS backup: Twilio + MessageBird + Plivo
  • Push notifications: Firebase FCM + Apple Push + OneSignal

​
Channel Failover

Set up multiple communication channels to ensure message delivery when specific channels fail or users lack contact information:

Channel Failover Example

Configure Best Of: Email → Push → SMS routing. Courier tries email first, then push if email fails, and finally SMS if both previous channels fail.
Configuration Options:
  • Template routing: Use “Best Of” channel configuration in template settings
  • Send API: Configure via message.routing property in your Send API requests
Strategic Benefits:
  • Contact coverage: Reach users even when primary contact info is missing
  • Delivery assurance: Multiple paths increase successful delivery rates
  • User preferences: Respect user channel preferences while maintaining backup options

​
Advanced Configuration

​
Timeout Management

Control when failover occurs by configuring timeout thresholds at different levels: Default Timeouts:
  • Provider timeout: 5 minutes (300000ms) - Time to wait for individual provider responses
  • Channel timeout: 30 minutes (1800000ms) - Time to attempt all providers in a channel
  • Message timeout: 72 hours (259200000ms) - Overall delivery attempt window
Timeout Hierarchy:
  1. Global timeouts: Apply to all providers/channels unless overridden
  2. Channel-specific timeouts: Override global settings for specific channels
  3. Provider-specific timeouts: Override global settings for specific providers
Precedence Rule: More specific timeouts take precedence. For example, message.providers.slack.timeout overrides message.timeout.provider for Slack delivery attempts.
Configuration Levels:
  1. Global Configuration:
  2. Channel-Specific Overrides:
  3. Provider-Specific Overrides:

​
Complete Configuration Example

This configuration creates a complete failover strategy with custom timeouts for time-sensitive notifications.