I am setting up a collector with Scala/Kinesis. First question:
Do I need a collector for each of my sites if they use a different domain? I am not able to find a solid answer on this on the GitHub getting started docs.
I am setting up a collector with Scala/Kinesis. First question:
Do I need a collector for each of my sites if they use a different domain? I am not able to find a solid answer on this on the GitHub getting started docs.
Please don’t ask multiple separate questions in the same thread - I’ve renamed your thread to your first question - please post a separate thread for your second question.
Do I need a collector for each of my sites if they use a different domain?
In general - no but this depends a lot on your use case, how many domains, what browsers you are targeting etc. A large number of users who use Snowplow to track usage across multiple domains will have a single collector that trackers across multiple domains reference.
There are some caveats to this as certain browsers restrict setting third party cookies for privacy reasons and Intelligent Track Protection (ITP 2.0) will also have some impact on this if you are planning on using third party identifiers (rather than just the Snowplow identifier from the first party cookie).
Sorry, it was literally my first post on here. I edited my post. Thanks for the tip.
@mike That makes sense. I think I’ll stick with first party cookies. Would there be issues if I used the same kinesis streams between 2 or more collectors?
edit: Can I cross analyze the data between two sites using first party cookies?
Although you could point two different collectors to the same Kinesis stream from a technical standpoint it’s not really a recommended implementation pattern (though again this depends on your use case) as it can conflate what domain the first party cookie has come from.
By the nature of first party cookies you can’t analyse them across sites (unless they are on the same site/subdomain). You’ll need some other identifier (e.g., a user_id or semi-persistent identifier) across different sites in order to link the same user together.