Sources are the origin of data. Each campaign in NBT must have a source where it gets data from.
They can be of different types, and more types will be included as new versions are released. Depending on the type of the source, data will be retrieved in different ways; you will have to choose depending on your license or your needs the one that suits best to each situation.
Current available are source types are nbtagent sources and manual sources.
Manual sources are available for all campaigns; these are the simplest ones, but cannot provide any kind of automation.
Data needs to be entered manually in the console, and output of each execution must also be retrieved manually.
When working with this type of source you have to prepare the input file yourself, and also you will have to workout through the ouput in order to feed your system.
These kind of sources are appropiate for campaigns that you will only execute once, or for any campaign during its testing phase.
Please note that using other type of sources will not prevent you from executing campaigns manually, so this type a source can be considedered as a default one that is included in every campaign.
nbtagent sources are available only with specific licensing (please contact email@example.com if you are interested).
They provide full automation and integration with your local repositories.
With this type of source you will need to install an agent in your system. This agent will act as a data collector from your data repositories, and also will update them whenever there are results from any execution.
From the console you will configure how the agent collects data (it can be from a file or it can connect to a database and read data from it), and how it updates the data (again, to a file or database).
A single nbtagent can act as a source for many campaigns, so you don´t need to install an agent for each campaign. Typically there will be just one agent per staging environment (production, preproduction), but if there are security/network concers there can be more than one.
You can also configure in the console the frequency that this data collection will be done, depending on your needs (daily, weekly, specific days ...), and at what time you want the collection to be executed.
Please note, that nbtagents are not daemon processes, so some scheduling needs to be configured in your systems,
As an example, a typical configuration will be to execute with this kind of scheduling.
- campaign A at "all labour days at 22:00"
- campaign B at "everyday at 16:00"
- campaign C at "monday, tuesday and thursday at 9:00"
You will configure each campaign collection process in the nbt console. But the agent needs to be executed to fit these scedulings, so our recomendation is to execute the agent each 10 minutes during the whole day.
You can try to acomodate the local scheduling of the agent to the campaign scheduling but then whenever you change it or add more campaigns, you will have to change both.
When the agent starts it will always ask the nbt server if it has anything to do, and most of the time will be told to "do nothing", so it will stop in just a few seconds.