What Happens When an EC2 Zonal Reserved Instance Expires?
When an Amazon EC2 Zonal Reserved Instance reaches the end of its one-year or three-year term, your running EC2 instance continues to operate without any interruption, reboot, or state change. Because a Reserved Instance is strictly a billing discount and capacity reservation construct rather than a physical virtual machine, its expiration only alters how AWS calculates your hourly compute charges and removes your Availability Zone capacity guarantee. Once the reservation expires, Amazon Web Services automatically transitions the billing rate for the matching running instance to standard On-Demand pricing unless another active reservation or Savings Plan applies to the usage.
Understanding the Separation Between Instances and Reservations
A common misconception in Amazon Web Services is treating a Reserved Instance as a distinct server. In reality, when you purchase an EC2 Zonal Reserved Instance, you are buying two things: a billing discount coupon and a regional Availability Zone capacity reservation that match specific attributes (instance type, platform, tenancy, and Availability Zone).
When you launch an EC2 instance that matches those attributes, the AWS billing engine automatically applies the discounted rate to that running resource. At the exact second the Zonal Reserved Instance term expires, nothing happens to the underlying hypervisor, network interfaces, Elastic Block Store volumes, or memory state of your running server. Your workloads remain completely online.
Immediate Billing Changes Upon Expiration
The primary impact of a Zonal Reserved Instance expiring while the matching EC2 instance continues to run is financial. The moment the reservation term ends, AWS evaluates your account for other applicable discount mechanisms before falling back to default rates:
- Other Active Reserved Instances: If you have another unallocated Regional or Zonal Reserved Instance matching the running instance attributes, it immediately takes over the discount coverage.
- AWS Savings Plans: If you have an active Compute Savings Plan or EC2 Instance Savings Plan with unused hourly commitment, that plan automatically covers the instance usage at its respective discounted rate.
- On-Demand Pricing: If no other commitment discounts are available, the running instance immediately reverts to standard On-Demand hourly or per-second billing rates.
To calculate the hourly cost increase when reverting to On-Demand pricing, let \(R_{\text{OD}}\) represent the On-Demand hourly rate and \(R_{\text{RI}}\) represent the effective hourly rate of the expired Zonal Reserved Instance. For \(N\) continuously running instances uncovered over \(H\) hours after expiration, the additional expenditure \(\Delta C\) is expressed as:
\[\Delta C = N \times H \times (R_{\text{OD}} - R_{\text{RI}})\]
If an instance runs for a full 730-hour month after expiration without renewal, the percentage increase in compute cost \(P_{\text{increase}}\) compared to the reserved rate is:
\[P_{\text{increase}} = \left( \frac{R_{\text{OD}} - R_{\text{RI}}}{R_{\text{RI}}} \right) \times 100\%\]
Loss of Availability Zone Capacity Reservation
Unlike Regional Reserved Instances, which provide instance size
flexibility across an entire AWS Region without reserving capacity, a
Zonal Reserved Instance is tied to a specific Availability Zone (such as
us-east-1a) and includes an On-Demand Capacity
Reservation.
While your instance is continuously running at the moment the Zonal Reserved Instance expires, the loss of the capacity reservation has no immediate operational effect because your instance already occupies its physical slot on the AWS host hardware. However, the capacity guarantee disappears immediately for future lifecycle events:
- Stopping and Starting the Instance: If you stop
your EC2 instance after the Zonal Reserved Instance has expired, you
release the underlying host hardware. When you attempt to start the
instance again later, AWS treats it as a new capacity request. If that
specific Availability Zone is experiencing a hardware shortage for that
instance family, you may encounter an
InsufficientInstanceCapacityerror. - Rebooting the Instance: Performing a standard instance reboot (either via the operating system or the AWS API) keeps the instance on the same physical host, so it is unaffected by the expired capacity reservation.
- Auto Scaling Replacement: If an Auto Scaling group terminates an unhealthy instance and launches a replacement in that Availability Zone after the reservation expires, the new launch is no longer guaranteed capacity.
Comparing State Before and After Expiration
| Attribute | Active Zonal Reserved Instance | Expired Zonal Reserved Instance |
|---|---|---|
| Running Instance Uptime | Continues running normally | Continues running without interruption |
| Hourly Billing Rate | Discounted effective rate (\(R_{\text{RI}}\)) | Reverts to On-Demand (\(R_{\text{OD}}\)) or active Savings Plan |
| AZ Capacity Guarantee | Reserved in the specified Availability Zone | No longer reserved; subject to AZ availability on stop/start |
| Instance Size Flexibility | None (locked to exact instance size) | Not applicable |
How to Prevent Unexpected Cost Spikes
To avoid unplanned On-Demand charges and capacity risks when a Zonal Reserved Instance approaches expiration, you can implement several preventative measures in AWS Cost Management and EC2:
- Configure Expiration Alerts: Use AWS Cost Explorer Reservation Expiration alerts to receive automated email notifications 7, 30, or 60 days before your Zonal Reserved Instances expire.
- Queue a Renewal Purchase: AWS allows you to queue the purchase of a replacement Reserved Instance up to three years in advance. By scheduling the queued purchase to trigger at the exact timestamp your current reservation expires, you maintain seamless discount and capacity coverage.
- Transition to Savings Plans plus ODCR: Many organizations replace expiring Zonal Reserved Instances by purchasing a Compute Savings Plan for flexible billing discounts while separately creating an explicit On-Demand Capacity Reservation (ODCR) for critical workloads that still require guaranteed Availability Zone capacity.