Hi guys!
As we didn't have any custom solution for Optimizely integration into Snowplow, we've upgraded Snowplow JS tracker to version 2.6.0 right after it's release and enabled Optimizely integration. Unfortunately, we were experiencing some issues with it (high latency), therefore we reverted JS tracker back to version 2.5.0.
Recently, I did some testing on Optimizely integration using various settings - see chart attached. It turned out that Optimizely integration parsing all data available causes quite high latency ~200ms. When I disabled Optimizely integration, but still using POST requests, latency dropped to ~30ms and when I switched to GET requests only, latency dropped to ~5ms. At the end I tried Optimizely integration (POST requests) parsing states data only and latency was ~50ms, but still quite noisy.
Clojure collector is running on two t2.medium instances, allowing to scale to up to four instances, but it never scaled up during a test.
I understand there is more data to be processed by the server when Optimizely integration is enabled, therefore I am wondering whether I should use larger EC2 instances or something else would do the trick? Is that something you experienced as well or did we do something wrong?
I appreciate any help you can provide.
Best,
Dejan