How does a retention policy impact a subscription?

I’m playing around with Kapacitor and having a hard time understanding how Retention Policies in InfluxDB have an impact on a stream. The stream seems to effectively be relaying data from InfluxDB to Kapacitor so Kapacitor can monitor the stream of incoming data and look for exceptions/out-of-bound conditions. I notice that retention policies feature prominently both in the subscription and in how Kapacitor binds to said subscription, so it makes me wonder - why does it matter?

If a retention policy specifies that data older than 1 year is removed from InfluxDB, how does that impact the real-time push of data out of InfluxDB?

Can retention policies on the Kapacitor InfluxDBsubscription improve or harm performance?

In Kapacitor both databases and retention policies are only used for scoping/namespacing the data. Meaning that it allows you to know where the data came from, specifically from which database and retention policy the data came from. Other than that they are just names to Kapacitor. The actual retention settings have no bearing on how Kapacitor or subscriptions behave.

Thanks Nathaniel, much appreciated. I was worried I was missing some core concept.