There was something wrong with fetching your inbox. Please contact support here.
Usage of the Haven OnDemand platform is packaged and controlled by the consumption of API Units and Resource Units.
API units are consumed when using API invocations. Resource Units are consumed when creating Haven OnDemand persistent resources, such as Text Indexes, Connectors, Query Profiles, and Prediction Models.
Consumption can occur by making programmatic API calls, or by using Haven OnDemand interfaces, such as Haven Search OnDemand and the API Explorer.
Information on the available Haven OnDemand APIs can be located here. Following are further details about the rates of consumption.
Each subscription plan includes a number of RUs. In standard usage, the number of RUs corresponds to an index size. For example, if you have a 10 GB standard flavor index, the index will cost 10 RUs, while a 1 GB categorization index will cost 15 RUs. Connectors also have different flavors, Onsite, and Cloud.
You can choose alternative index configurations, such as Query Manipulation, and Categorization. Using different index configurations can result in a smaller or greater index size. For more details on resource units and text index resource types, click here.
All API calls consume API units in accordance with the table provided here. Typically, invoking one API consumes 1 API unit. Each subscription plan additionally includes a set number of search API calls (those that accept an index parameter) against user-created text indexes.
Calling combination APIs consumes multiple API units, because combinations are aggregates of APIs. The number of API units corresponds to the APIs that the combination uses.
Further, some APIs incur more than one API unit whenever you call them. For example, our Speech recognition API uses 1 API unit when you call it, and then consumes 1 API unit for each second of speech. So, every minute of speech recognition consumes 60 API units.
Haven OnDemand is designed as a high availability platform, capable of responding to multiple API requests concurrently. HPE uses automated technology to measure platform availability every ﬁve minutes. Haven OnDemand defines platform uptime as the duration of time when the majority of APIs can successfully handle a predefined request. For any calendar month in which platform availability was less than 99.95%, HPE provides a credit on your invoice to the value of 20% of the monthly total.
Should you exceed the RUs or API units provided with your subscription package, additional units may be consumed at the overage rate. Haven OnDemand applies the overage rate unit price for each additional 10,000 API units or 1 resource unit.
To prevent excess overage, Haven OnDemand sets your quota to twice the selected size of your subscription plan. If you require a higher quota, please contact us at firstname.lastname@example.org. For additional details on quotas, click here.
Based on your billing location, Haven OnDemand adds sales and other mandatory taxes during your subscription checkout.
Haven OnDemand may archive resources created under our Freemium plan after a period of inactivity.
Lower rate limits may be applied for certain API types and a number of factors may impact actual API throughput. Rate limits are not guaranteed throughput rates - further details are available here.