- 최신
- 최다 투표
- 가장 많은 댓글
The common workaround for launches—re-sending the request without specifying AZ—cannot be used to restart stopped instances. Effectively, there are limited workarounds available.
1. Temporarily change the instance type (most practical approach)
This is the most commonly reported practical workaround. Simply switching to a type with similar specifications within the same generation and family (e.g., using m5a.large or m6i.large if m5.large is unavailable) is often enough to avoid capacity issues. You can revert to the original instance type later.
2. Create an On-Demand Capacity Reservation
This mechanism allows you to reserve capacity for a specific AZ and instance type in advance. For critical production instances, this is worth considering as a permanent solution.
When you encounter an insufficient capacity error while starting a stopped EC2 instance, it means AWS doesn't currently have enough available On-Demand capacity in that specific Availability Zone to fulfill your request. While waiting and retrying is one option, there are several other approaches you can take:
Immediate alternatives:
- Create a new instance from a snapshot in a different Availability Zone - This allows you to move your workload to an AZ with available capacity
- Try a different instance type - Submit a new request using a different instance type that may have better availability
- Launch without specifying an Availability Zone - Let AWS automatically select an AZ with available capacity
Proactive solutions:
- Configure alternate instance types - If you're using Instance Scheduler or similar automation, you can configure a prioritized list of alternate instance types. The system will automatically attempt to resize your instance to alternative types (that are compatible with your configuration) if the preferred type encounters capacity issues
- Reduce the number of instances - If you're starting multiple instances simultaneously, try starting them in smaller batches
Important considerations:
- Any alternate instance type must be compatible with your instance's current configuration (AMI, storage, networking, etc.)
- Capacity constraints are typically temporary and vary by region, AZ, and instance type
- Different instance types and AZs may have different availability at any given time
These approaches can help improve your instance availability in capacity-constrained situations beyond simply waiting and retrying.
Sources
Troubleshoot Amazon EC2 instance launch issues - Amazon Elastic Compute Cloud
Resolve insufficient instance capacity errors in Lightsail - Amazon Lightsail
Handling EC2 Insufficient Capacity Errors - Instance Scheduler on AWS
