Scheduling campaigns only apply to campaigns configured with a source of type nbtagent.
Other types of source involve a manual process, and hence it doesn't make sense to schedule them.
Here you will be able to schedule when the nbtagent will execute the data collection in order to begin an analysis.
Usually nbtagents will be executing many times each day and will ask the nbt server if they have any action pending.
If the server detects that is time for the agent to collect data for a specific campaign, it will send the appropiate message and the nbt agent will perform the data collection.
During this process the server will use the information configured here, to tell the agent that it is time to execute or just to don't do anything for this campaign.
As an example, let's look at this configuration :
Scheduling for this campaign is set to :
labour days (from monday to friday)
- at 22:00
With this settings if the agent executes at :
- tuesday at 21:10. The server will tell the agent to don't do anyting
tuesday at 21:40. The server will tell the agent to don't do anyting
- tuesday at 22:10. The server will tell the agent to collect data, as it is more than 22:00.
- tuesday at 22:40. The server will tell the agent that it has data pending from a previous execution, and it will be told to download and process it.
- tuesday at 23:10. The server will tell the agent to don't do anyting.
- saturday at 21:10. The server will tell the agent to don't do anyting
saturday at 21:40. The server will tell the agent to don't do anyting
saturday at 22:10. The server will tell the agent to don't do anyting.(saturday is not a labor day, so the agent won't be told to do anything.
- saturday at 22:40. The server will tell the agent to don't do anyting.
- saturday at 23:10. The server will tell the agent to don't do anyting.
Using this panel you can configure how often you want the agent to collect the data (everyday, labour days or specific days of the week), and at what time.
Developmente mode should be used when you are configuring a campaign. Setting a campaign to use this mode, will make the agent begin the collection process every time that it is executed.
It is appropiate for troubleshooting or when configuring a campaign for the first time, so you can quickly check that it will work fine.
Using this setting you can troubleshoot problems with
- the sql statement,
- data types on your model,
- problems that may arise with the content of specific columns in your model because of certain features of your data (unique values or too low variation in some columns)
including dates in your models, which is not appropiate as they are usually are not good predcitors.
It is recommended when using this approach that you can manually execute the campaign, instead of having to wait for the agent to be executed. The reason is to have an agile process of executing, checking errors, correct them, execute again, and so on.
This setting will only last for an hour when set, to prevent the situation where a campaign is left in this mode for a long period of time. If you need more time, you can come here at anytime and set it again.
It is important to note, that in Development mode, only the data collection and the analysis will be executed.
Updating scores will not be done when in Development mode. However you can always go to the activitites section and download the ouput of any execution, including the Development mode ones.
Executions in Development mode are marked with a (D) in the activities section.
See also :