Skip to content

Getting an instance

Before asking for an instance make sure your use case cannot be satisfied with the global CERN Search.

To obtain your instance, please open a request through Service Now, and be sure to provide the following information:

  • Project information: A short description of the project where the instance will be used and the use case, by whom, and who (Experiment, Department, Section or Service) will be the owner and therefore responsible of its usage.

  • Access permissions: Egroup responsible for the instance and therefore with administrative access to it. This egroup is not trivial to be changed, please choose it carefully.

  • Superuser permissions: Email/username as returned by SSO of the account that will have superuser permissions.

  • Desired URL: The name of the desired URL for your endpoint. We will do everything in our hands to fulfill your requirements, if not possible we will adapt it to be as similar as possible. The naming of the instances is constrained by the Webservices policies.

  • Instance Specifications:

    • Amount of collections.
    • Amount of documents (number of document to index, per collection).
    • Size estimation of the total space your documents might take.


    This will enable us to help you designing the documents' mappings and schemas to achieve optimal performance. If you have doubts about what is a document, a collection or others please refer to the concepts section.

  • Document(s) description: A description of the attributes of each document, specifying for each one:

    • Type: Numeric, date (And format, e.g. yyyy-MM-dd'T'HH:mm:ss), string, etc.
    • Type of search: Exact match or full text search.
    • Type of access: If the field needs to be searchable through a global search or will only be used for refiners or filters (e.g. By file type, content source).
    • Storage only: If the content of the attribute must be returned in the response, but it does not have to be indexed for search (e.g. some of your internal user ID/PK that will help you to perform actions in your application but you will not query the search endpoint for it).
    • Languages: Languages that the attribute can be in (This will help with features like stemming). By default they are processed for English and French.
    • Binary content: If the document is a binary file, we will discuss the best way to extract the content. You can see more in binary text content extraction.