Collector Clojure vs Cloudfront

[Joao Correia]

Hi Snowplowers,

I've been using Cloudfront and it works fine, why would I want to change to Clojure? Aren't the thrid party cookies blocked more often than first party? 

I would like to hear other's experiences on the collectors. 

Thanks

Joao

[Ihor Tomilenko]

Hi Joao,

If Cloudfront collector satisfies your business requirement then by all means stick to it. Cloudfront is simple and reliable, it's powered by Amason's CDN. If you are not interested in cross-domain tracking and are not implementing complex custom events and contexts, Cloudfront collector should serve you well. 

One might find the following disadvantages though. Cloudfront collector:

1. Does not support POST requests (only GET)
   
2. Logs could be delayed by more than 1 hour, sometimes up to 24 hours (according to Amazon).
   
3. And as mentioned before, it does not support cross-domain tracking
   

The first item is important when you send custom events/contexts. The GET request is limited by 2048 characters. Thus if your tracker sends a large amount of data then it could be an issue.

Clojure collector address this limitation by 

1. Allowing to send data via POST (theoretically unlimited size) in addition to GET. 
   
2. The logs are provided by Tomcat and rotated to S3 bucket on hourly bases.
   
3. Sets cookies server-side thus enabling cross-domain tracking (3-rd party cookies).
   

Clojure collector is also scalable but by means of Amazon's Elastic Beanstalk (as opposed to CDN) technology.

You might also be interested in reading this blog article devoted to Clojure collector.

Hopefully this answers your question.

Regards,

Ihor