Pilot Light Pattern
What You'll Learn
You'll learn how to maintain a minimal but critical infrastructure footprint in your disaster recovery region that can be rapidly expanded when disaster strikes.
The Core Concept
The Pilot Light pattern takes its name from the small, always-on flame in a gas furnace that's ready to ignite the main burners instantly. In disaster recovery, you keep only the most critical core components running continuously in your DR region—typically databases with live replication and perhaps minimal compute instances—while leaving everything else switched off.
When disaster occurs, you "turn up the heat" by launching additional application servers, scaling out compute resources, and activating dormant services. This approach strikes a balance between the zero-infrastructure Backup and Restore pattern and the fully replicated Warm Standby or Hot Standby patterns.
What Stays On vs. What Launches
Always Running (The Pilot Light):
- Database replicas with continuous replication
- Core configuration management systems
- Critical security and networking infrastructure
- Essential monitoring agents
Launched During Failover:
- Application server fleets
- Load balancers and auto-scaling groups
- Caching layers and message queues
- CDN configurations and DNS failover
Why This Pattern Works
The pilot light pattern significantly reduces DR costs compared to maintaining fully active infrastructure in multiple regions. You're only paying for minimal compute and storage replication, yet you can achieve RTOs measured in minutes to hours rather than days. The trade-off is that recovery requires automation scripts and tested runbooks to orchestrate the rapid scale-up.
Key Takeaway: The Pilot Light pattern maintains just enough critical infrastructure in your DR region to enable rapid expansion when needed, balancing cost efficiency with reasonably fast recovery objectives.