Does Amazon EC2 Guarantee a Regional Monthly Uptime SLA?

Amazon Elastic Compute Cloud offers a formal Service Level Agreement that guarantees a specific monthly uptime percentage for resources deployed across an AWS Region. Under the Amazon Compute Service Level Agreement, AWS commits to making Amazon EC2 available with a Monthly Uptime Percentage of at least 99.99 percent during any monthly billing cycle when instances are deployed across multiple Availability Zones within the same region. For single EC2 instances, AWS also provides an instance-level availability commitment of 99.5 percent. When these availability targets are not met, eligible customers can claim financial service credits applied against future Amazon EC2 payments.

Understanding the Regional-Level SLA Commitment

The regional-level SLA for Amazon EC2 is designed around high-availability architectures that span more than one Availability Zone. An Availability Zone consists of one or more discrete data centers with redundant power, networking, and connectivity inside an AWS Region.

To qualify for the 99.99 percent regional Monthly Uptime Percentage commitment, your workload must run concurrently across two or more Availability Zones in the same region. A region is considered unavailable—referred to in the agreement as Region Unavailable—when all of your running instances deployed across two or more Availability Zones concurrently have no external connectivity. If only one Availability Zone experiences an outage while another Availability Zone containing your running instances remains healthy and reachable, the regional SLA threshold is not breached.

Single Instance SLA Commitment

In addition to the multi-zone regional commitment, Amazon EC2 includes an Instance-Level SLA for individual virtual machines. AWS commits to a Monthly Uptime Percentage of at least 99.5 percent for any single EC2 instance during a monthly billing cycle. A single instance is considered unavailable when it has no external connectivity and fails to respond to external network requests due to underlying AWS infrastructure issues.

How Monthly Uptime Percentage Is Calculated

AWS calculates the Monthly Uptime Percentage by subtracting from 100 percent the percentage of minutes during the monthly billing cycle in which the service was in a state of unavailability. The mathematical formula used for the regional calculation is:

\[\text{Monthly Uptime Percentage} = 100\% - \left( \frac{\text{Total Unavailable Minutes in a Billing Cycle}}{\text{Total Minutes in a Billing Cycle}} \times 100\% \right)\]

If an instance has only been running for part of the month, AWS assumes the instance was 100 percent available for the portion of the month it was not running. Furthermore, any downtime lasting less than one full minute is generally not recorded as part of the unavailable minutes unless specified by the monitoring thresholds of the agreement.

Service Credits for SLA Breaches

When the Monthly Uptime Percentage falls below the guaranteed threshold, AWS compensates customers through Service Credits. The credit percentage scales based on the severity of the downtime experienced during the billing cycle.

For the Regional-Level SLA across multiple Availability Zones, the credit tiers are structured as follows:

  • Less than 99.99 percent but equal to or greater than 99.0 percent results in a 10 percent Service Credit.
  • Less than 99.0 percent but equal to or greater than 95.0 percent results in a 30 percent Service Credit.
  • Less than 95.0 percent results in a 100 percent Service Credit.

For the Single EC2 Instance SLA, the credit tiers are:

  • Less than 99.5 percent but equal to or greater than 99.0 percent results in a 10 percent Service Credit.
  • Less than 99.0 percent but equal to or greater than 95.0 percent results in a 30 percent Service Credit.
  • Less than 95.0 percent results in a 100 percent Service Credit.

Service credits are applied only against future Amazon EC2 charges for the affected region or instance and are not paid out as direct refunds.

SLA Exclusions and Limitations

Not all service disruptions count toward SLA downtime calculations. The Amazon Compute SLA outlines several exclusions where unavailability does not qualify for service credits:

  • Factors outside of AWS reasonable control, including force majeure events or broad internet access problems beyond the demarcation point of the AWS network.
  • Actions or inactions taken by the customer or third parties, such as misconfigured security groups, network access control lists, routing tables, or operating system kernel panics.
  • Failures resulting from customer equipment, third-party software, or custom scripts.
  • Downtime occurring during scheduled maintenance windows where AWS has provided advance notice, or instances terminated due to breaches of the AWS Acceptable Use Policy.
  • Spot Instances, which are designed to be interruptible when AWS needs to reclaim spare compute capacity.

Requirements for Claiming a Service Credit

Service credits are not issued automatically in most scenarios. To receive a credit for a breach of the Amazon EC2 SLA, an account administrator must submit a formal claim by opening a case in the AWS Support Center before the end of the second billing cycle after the incident occurred. The request must include the specific dates, times, and affected Region or Instance IDs, alongside request logs that document the errors and corroborate the claimed outage duration.