How does the Alerts procedure work?

• Article Type: General
• Product: MetaLib
• Product Version: 3

Description:
How does the Alerts procedure work?

Resolution:
The Alerts process is generated by UTIL K-1, which is run on the server. The procedure checks all saved alerts, and based on the last run date and the interval for the run, decides whether or not to run each alert.
By default, the system sends 5 alerts at a time. Every alert is a query run against up to the maximum number of simultaneously searched resources. After 50 seconds, if results do not come in, the search is suspended. Most alerts will take less than 50 seconds.
The alerts use the same mechanism as that used for online searches. This means that when the alerts procedure is run, it is as if there are 5 users online running 5 queries at the same time on one or more resources each. The alerts are sorted by the User ID and a running sequence. Note that only alerts whose predefined interval has passed will be run. If a user has more than one alert that should be run on the same day, these will most probably run in the same group of 5.
It is possible to change the number of alerts that are activated at a time. This can be done when the alerts are run via a cron job.
A site that has resources with very limited licenses (e.g., 1-2) may want to run the alerts one at a time. This will prevent the alerts from failing due to a license limit. (A similar problem would occur online if several users attempted to search the resource at the same time). If you make such a change, the alerts will take longer to run. This should not present a problem, unless you have a very large number of alerts running per day. You can monitor the amount of time the alerts procedure takes in the log file created.