In cloud services, whether it is infrastructure as a service (IaaS), platform as a service (PaaS), or software as a service (SaaS), service availability is often a significant customer concern because the customer is relying on the vendor to provide and manage the infrastructure and related components that are necessary to provide the services. To address this concern, vendors will often provide a Service Level Agreement (SLA) containing a commitment that the service will be available for a percentage of time (e.g., 99.9%) during a certain period (e.g., week, month, or quarter). This is often referred to as an uptime or availability commitment. When reviewing and negotiating an SLA with an uptime commitment, it is important to consider the following issues.
Given the different types of cloud services and how those services are used, there is no standard uptime commitment provided by vendors. Rather, uptime commitments can range from 99.999% to 97% or even lower. It is also not uncommon for vendors to provide different uptime commitments for different parts of the service. Ultimately, a vendor’s uptime commitment will depend on a variety of factors, including the type of service, how a customer will use the service, negotiating leverage, and vendor’s business model.
The period during which the uptime commitment is calculated is important. For example, an uptime commitment of 99.9% each month means there could be about 43 minutes of downtime in a particular month; however, if the uptime commitment is calculated every quarter then downtime could exceed two hours, which could significantly impact the customer and its business. Generally, shorter periods for calculating the uptime are better for customers, while vendors will want a longer period.
Customers should seek an uptime commitment that is appropriate in relation to the type of service and how the service will be deployed. For example, if a service is mission critical where downtime would have a significant impact on a customer’s revenue then the customer should push for an uptime percentage with multiple nines. Conversely, if a service is ancillary to a customer’s business and any downtime impact would be negligible, a customer may be fine with whatever commitment the vendor offers.
Vendors may tout their services as having an excellent uptime commitment (e.g., 99.99%); however, it’s important to understand the details regarding the uptime commitment. Notably, vendors will often seek to exclude numerous events from the calculation of the uptime commitment. At minimum, the exclusions will likely include planned downtime; however, vendors will likely seek to carve out other events from the uptime calculation, including issues outside of the vendor’s control (such as force majeure events) and issues caused by customers.
While it is reasonable for the vendor to exclude certain events from an uptime commitment, the scope of these exclusions should not be so broad as to exclude events that are actually within the vendor’s control (e.g., issues caused by vendor’s suppliers.) If the exclusions are unreasonably broad, the uptime commitment offered by the vendor may not provide any meaningful protection.
The most common remedy for a vendor missing its uptime commitment is a refund or credit against future invoices. The refund or credit will typically be structured as a percentage of the fees paid over a certain period (e.g., month or quarter). For example, a vendor may state that a customer is entitled to a fixed percentage credit or refund for each percentage that the vendor is below its uptime commitment (e.g., 1% of fees for every 0.5% below the uptime commitment). Typically, the refund or credit will be capped at a fraction of the fees paid during the applicable measurement period (e.g., 10% or 15%).
In addition to or in lieu of a refund or credit, a vendor may offer the right to terminate the agreement if there is a persistent failure to meet the uptime commitment or if the refund or credit reach the cap. For example, a vendor may allow a customer to terminate its agreement if the vendor fails to meet the uptime commitment in three consecutive months or in any four months within a 12-month period. Importantly, if a customer is allowed to terminate, the customer should seek to at least receive a refund of any unused, prepaid fees covering the remainder of the agreement.
Often, an SLA will state that the service credit or refund is the customer’s sole and exclusive remedy for vendor’s failure to meet its uptime commitment. While the vendor is seeking to minimize its exposure, the limited credit or refund likely will not make the customer whole if there is a significant downtime issue. Therefore, the customer should seek to remove this language or to make sure that the sole and exclusive remedy is narrowly tailored, so that the customer does not broadly waive its right to other claims under the agreement.
Tracking Uptime and Reporting
Vendors may provide a readily accessible mechanism for customers to monitor performance against the uptime commitment. This information may be publically available on a vendor’s website. Reviewing a vendor’s historical performance can help when trying to negotiate better SLA terms or deciding whether it is worth requesting changes to the SLA.
As a condition to receiving a credit or refund, vendors often require that customers report a service availability issue within a certain time period, which may range from a few days to a month. What is more, vendors may impose specific requirements regarding how a request for a credit or refund must be submitted (e.g., through vendor’s portal or to a specific email address). If specific requirements are imposed, customers must seek to implement an internal mechanism to monitor the vendor’s uptime commitment and to report issues to the vendor.