Working with a client and have recommended them to get Snowplow Collector, due to various requirements the Clojure collector is the best option, however being a large enterprise they will not sign off Elastic Beanstalk and will only load the Clojure collector directly to a EC2.
So currently in the process starting a POC to make sure this could work, however would appreciate some guidance on how to configure the logging part to make sure right data is captured.
Hi @fwahlqvist - this won’t work I’m afraid, the Clojure Collector is dependent on running inside a customised Apache Tomcat environment, with Beanstalk handling the log rotation.
The Scala Stream Collector is the one to go for if you want to run on EC2.
for the Scala Stream collector. Although it’d be possible to emulate the Clojure collector on EC2 you’d have to write new existing logic around handling the custom log rotation to S3 - building and testing this would be a fair amount of effort.
Out of curiosity what’s the reason for the client not being happy to sign off on Beanstalk?
@alex, Getting the collector working on EC2 was pretty straight forward as a POC, still need to adjust for clients AMI but don’t think this will be a issue. When is comes to log rotation we are still doing some testing but currently looks ok.
@Mike, would have loved the Scala stream collector but due to many different factors this was not a option either, Beanstalk was not allowed as it don’t conform to their CI/CD process.
As mentioned thanks for feedback!