External Source URL

This feature is part of release 23.3
Major update in release 24.3 offering asynchronous validation for data objects

Introduction

Instead of providing a file directly at the time of ingest, an external source URL is provided instead. The system downloads the file from the URL during the ingest.

Features

The feature is available in the API: https://integration.mediahaven.com/mediahaven-rest-api/v2/api-docs/index.html#url_file_uploadusing the parameter fileUrl when POSTing the record
The URL is stored in the read-only metadata field Internal.ExternalSourceUrl
If the MD5 checksum, calculated after downloading the URL, matches with an existing record, the new record will be rejected

Protocols

The only supported protocols are HTTPS and HTTP
Future versions might extend this feature to include object stores such S3
For now, for object stores create a temporary URL for the resource or make the object store public and access it via HTTPS.

Business Rules

General rules

The URL must adhere to https://www.rfc-editor.org/rfc/rfc1630 in particular unsafe characters such as spaces, control characters, some characters whose ASCII code is used differently in different national character variant 7-bit sets, and all 8-bit characters beyond DEL (7F hex) of the ISO Latin-1 set, shall not be used unencoded
Redirects are not allowed

Synchronous validation when POSTing the record

Up to version 24.2 the external URL is always validated synchronously meaning that the URL is validated syntactically. An HTTP HEAD request is done to obtain the file size using the HTTP response header “Content-Length”. From version 24.3 for Data Objects the validation is done asynchronously and the HTTP header Content-Length is no longer required.

Rule	Up to version `24.2`	From version `24.3` for flat data objects	From version `24.3` for Data Objects

Rule	Up to version `24.2`	From version `24.3` for flat data objects	From version `24.3` for Data Objects
HEAD request returns `200 OK`	Yes	Yes	No, it is checked asynchronously during ingest When the validation fails the record will be rejected
The header `Content-Length` is present in the HEAD response and is a valid positive number	Yes	Yes	No longer validated or mandatory because the file size is determined by actually downloading the file during ingest

Asynchronous validation during ingest

The following rules apply to the external source URL when the system internally downloads the URL

The URL responds to a GET with a valid response with status 200 OK
Responses different from 200 OK will be retried with a pause of either 10 seconds, or up to 4 additional times, for a total of 5 attempts
- When the response is 429 Too many requests and it contains the response header Retry-After the system will pause with this amount of time instead
Redirects are not allowed