In cases where it is not possible for a remote system to send us the data, we can make a scheduled request for it using Pollers. They must, therefore, obviously belong to an integration. Although a Poller belongs to only one integration, an integration can have multiple Pollers.
A Poller is a configuration record which defines the frequency of polling and which logic to use (the logic itself is defined in the Poll Processor). Each time it is run, it creates a corresponding Poll Request record.
Depending on the use case, the use of a Poller to collect data from a remote system poses some development challenges which need to be considered. Namely, there is an additional responsibility and workload placed on the host system to store and check some returned data in order to evaluate what has changed.
For example, in order to decide whether the state has changed, or what comments have been added, or even which system has made the updates to the data (we don’t want to pull back data we have changed), checks have to be built into the scripts. This is aided by holding a copy of the relevant returned data, using Data Stores (more on that in the Administration section).