Vertex AI Platform Service Level Agreement (SLA)

During the Term of the agreement under which Google has agreed to provide Google Cloud Platform to Customer (as applicable, the "Agreement"), the Covered Service will provide a Monthly Uptime Percentage to Customer as follows (the "Service Level Objective" or "SLO"):

Covered Service

Monthly Uptime Percentage

Training, Deployment, and Batch Prediction

>= 99.9%

AutoML Tabular and AutoML Image Online Prediction for models deployed on 2 or more nodes

>= 99.9%

AutoML Text Language Online Prediction

>= 99.9%

Custom Model Online Prediction for models deployed on 2 or more nodes

>= 99.5%

Vertex Pipelines

>= 99.5%

Covered Service

Monthly Uptime Percentage

Training, Deployment, and Batch Prediction

>= 99.9%

AutoML Tabular and AutoML Image Online Prediction for models deployed on 2 or more nodes

>= 99.9%

AutoML Text Language Online Prediction

>= 99.9%

Custom Model Online Prediction for models deployed on 2 or more nodes

>= 99.5%

Vertex Pipelines

>= 99.5%

If Google does not meet the SLO, and if Customer meets its obligations under this SLA, Customer will be eligible to receive the Financial Credits described below. Monthly Uptime Percentage and Financial Credit are determined on a calendar month basis per Project. This SLA states Customer's sole and exclusive remedy for any failure by Google to meet the SLO. Capitalized terms used in this SLA, but not defined in this SLA, have the meaning stated in the Agreement. If the Agreement authorizes the resale or supply of Google Cloud Platform under a Google Cloud partner or reseller program, then all references to Customer in this SLA mean Partner or Reseller (as applicable), and any Financial Credit(s) will only apply for impacted Partner or Reseller order(s) under the Agreement.

Definitions

The following definitions apply to the SLA:

  • "Back-off Requirements" means, when an error occurs, Customer is responsible for waiting for a period of time before issuing another request. This means that after the first error, there is a minimum back-off interval of 1 second and for each consecutive error, the back-off interval increases exponentially up to 32 seconds.
  • "Covered Service" means the applicable Vertex AI Platform service listed under the 'Covered Service' column in the table above.
  • "Downtime" means more than a 5% Error Rate. Downtime is measured based on server side Error Rate.
  • "Downtime Period" means a period of one or more consecutive minutes of Downtime. Partial minutes or intermittent Downtime for a period of less than one minute will not count towards any Downtime Periods.
  • "Error Rate" means the number of Valid Requests that result in a response with HTTP Status 500 or 503 divided by the total number of Valid Requests during that period. Repeated identical requests do not count towards the Error Rate unless they conform to the Back-off Requirements.
  • "Financial Credit" means the following for Training, Deployment, and Batch Prediction; AutoML Tabular and AutoML Image Online Prediction for models deployed on 2 or more nodes; or AutoML Text Online Prediction:

Monthly Uptime Percentage

Percentage of monthly bill for the respective Covered Service that does not meet SLO that will be credited to Customer's future monthly bills

99% to < 99.9%

10%

95% to < 99%

25%

< 95%

50%

Monthly Uptime Percentage

Percentage of monthly bill for the respective Covered Service that does not meet SLO that will be credited to Customer's future monthly bills

99% to < 99.9%

10%

95% to < 99%

25%

< 95%

50%

  • "Financial Credit" means the following for Custom Model Online Prediction for models deployed on 2 or more nodes, or Vertex Pipelines:

Monthly Uptime Percentage

Percentage of monthly bill for the respective Covered Service that does not meet SLO that will be credited to Customer's future monthly bills

99% to < 99.5%

10%

95% to < 99%

25%

< 95%

50%

Monthly Uptime Percentage

Percentage of monthly bill for the respective Covered Service that does not meet SLO that will be credited to Customer's future monthly bills

99% to < 99.5%

10%

95% to < 99%

25%

< 95%

50%

  • "Monthly Uptime Percentage" means total number of minutes in a month, minus the number of minutes of Downtime suffered from all Downtime Periods in a month, divided by the total number of minutes in a month.
  • "Valid Requests" are requests that conform to the Documentation, and that would normally result in a non-error response.

Customer Must Request Financial Credit

In order to receive any of the Financial Credits described above, Customer must notify Google technical support within 30 days from the time Customer becomes eligible to receive a Financial Credit. Customer must also provide Google with identifying information (e.g., project ID and job IDs) and the date and time errors occurred. If Customer does not comply with these requirements, Customer will forfeit its right to receive a Financial Credit.

Maximum Financial Credit

The total maximum number of Financial Credits to be issued by Google to Customer for any and all Downtime Periods that occur in a single billing month will not exceed 50% of the amount due by Customer for the Covered Service for the applicable month. Financial Credits will be made in the form of a monetary credit applied to future use of the Service and will be applied within 60 days after the Financial Credit was requested.

SLA Exclusions

The SLA does not apply to any (a) features designated pre-general availability (unless otherwise stated in the associated Documentation); (b) features excluded from the SLA (in the associated Documentation); or (c) errors (i) caused by factors outside of Google's reasonable control; (ii) that resulted from Customer's software or hardware or third party software or hardware, or both; (iii) that resulted from abuses or other behaviors that violate the Agreement; (iv) that resulted from quotas applied by the system or listed in the Admin Console; or (v) that resulted from Customer use of the Covered Service inconsistent with the Documentation, including but not limited to invalid request fields, unauthorized users, inaccessible data, or use of a model that is beyond the recommended model lifespan described in the applicable Documentation.

Previous versions (Zuletzt geändert: 7. Juni 2023)
Google Cloud
  • ‪English‬
  • ‪Deutsch‬
  • ‪Español‬
  • ‪Español (Latinoamérica)‬
  • ‪Français‬
  • ‪Indonesia‬
  • ‪Italiano‬
  • ‪Português (Brasil)‬
  • ‪简体中文‬
  • ‪繁體中文‬
  • ‪日本語‬
  • ‪한국어‬
Console
Google Cloud