In this article you’ll find:
| Table of Contents |
|---|
Goal
The goal of this article is to show you how to create additional data sources which are a part of the indexing process as well as how the system handles data source failures.
Prerequisites
| Info |
|---|
|
Overview
You can add additional data sources for specific indexes which become a part of the indexing process. When the system begins a full indexing scheduled task, it sends a post request to each data source with a predefined API Key as well as the new Hawksearch index name so they can begin indexing as well. Those data sources are then responsible for sending status update requests back which determine if indexing fails or not based on settings. Explained in more detail below.
Adding additional data sources
Open the side menu application selector (F2 or button top left) and go to Custom / Hawksearch.
Go to Web Api / Data Sources.
Click New data source.
Name - choose a unique name for this data source.
API Key - choose a unique random API Key which will be used for this data source. This will be sent to the data source to ensure the validity of the request so we recommend choosing a random GUID.
Index name - choose the index for which you wish this data source to be. Predefined names coming from Kentico: Manage indexes .
Url - the url to which the system will send a post request along with the API Key and the new Hawksearch index name.
Severity - you can choose between Low & Critical. There are differences in the way the system handles failure based on this value. Explained further below.
Timeout interval - choose a duration in minutes. This duration determines how long the system will wait after notifying the data source before it is marked at timed out.
Workflow
When a full indexing scheduled task runs, the system notifies external data sources by sending the API Key for that data source as well the new Hawksearch index name. Each data source is then responsible for sending status update requests back to the system. If a data source doesn’t respond for a period longer than Timeout interval (in minutes), it is marked as timed out. If a data source times out or has a failing status update, indexing can fail based on the severity of the data source.
...