While it is true that in the large environments we usually wish to monitor for problems, in small environments we might wish to be inventive and also monitor for positive, interesting events.
Consider the following two scenarios:
(1) Suppose we have a self-made bank account. When financial resources come in, a log entry is written to a file. So whenever that happens, it is good news and we wish to be alerted.
(2) Suppose we run a small advertisement business and we have a huge printer. It takes a lot of time to print a poster, maybe on the order of hours. Using SNMP, we wish to monitor the printer for activity and be alerted when it has finished printing, so that we can immediately start a new printing task and we do not have to check the printer ourselves every now and then.
It is possible to receive alerts on such events using everything we have in Zabbix right now. Namely, we would create triggers and define alerts. The problem is that the only way to receive repeated alerts based on triggers is to have triggers in the PROBLEM state, but in neither of the two scenarios above the events we wish to monitor for are problems. We could use the existing mechanisms, but the "good" PROBLEM triggers would interfere wish "real problem" PROBLEM triggers.
Scenario (1) can be solved, for instance, by either (a) introducing an ability to receive alerts whenever a new value comes in or (b) allow triggers to generate multiple OK events.