Rich-text Reply

Experiencing very poor performance with Optimizely Event Log

kmikhai 12-24-16

Experiencing very poor performance with Optimizely Event Log

Hi. I'm relatively new to Optimizely. 

 

I am a lead of technical team, and we're supporting an app that uses optimizely experiments and event logging. We have very strong instrumentation for the app, and so we're noticing that our optimizely event logging calls specifically tend to experience very poor performance, with a good portion of calls (third to a half) taking over 1 second, and quarter taking 3 seconds or more, sixth of them taking 7 seconds or more per call (!).

 

In further analysis of the event log dependency, I noticed that our optimizely event logs happen to a dedicated AWS unit, using our optimizely tag as domain prefix.

 

The poor and uneven performance of optimizely event log is significantly impacting our conversion rates, and of course that's counter productive since we're using Optimizely with intention of improving site performance, not the opposite.

 

Can you please help with some ideas - perhaps the event log instance can be scaled? We have a paid support account, but since I am new to this I am not on it, though I can get all info needed if this becomes a true support case.

Level 1

tedroddy 12-28-16
 

Re: Experiencing very poor performance with Optimizely Event Log

Hi there,

The events should be sent asynchronously in the background, so your users would not experience a slowdown or block during the event dispatch. 7 seconds is very long indeed, but what is the mean latency in your data? Those could just be outliers, and given that many users have poor connectivity in mobile apps it wouldn't be out of the question to see longer calls like this.
Best,
Ted
Optimizely
cubelodyte 02-08-17
 

Re: Experiencing very poor performance with Optimizely Event Log

[ Edited ]

We appear to be having similar issues on a sight that is currently not even running an active experiment. 

I'd love to hear how this was resolved. 

 

The attached image shows our analysis tool with a single event "blocking" at 1.56s. This is an irregular but frequent occurance on this site and none of the experiments (Optimizely Classic) are active. 

 

If it helps, we've recently converted parts of this site to react.js. Prior to this migration, we were not seeing this issue. Screen Shot 2017-02-07 at 8.32.41 AM.png

Scott Ehly
Manager of Site Optimization
sehly@rentpath.com

'The single biggest problem with communication is the illusion that it has taken place.' - George Bernard Shaw
juliofarfan 02-08-17
 

Re: Experiencing very poor performance with Optimizely Event Log

Hello! We've experienced the same "trouble" that appears even in console, it doesnt matter if experiment is active or not and I believe that its related with the time the experiment has been on your dashboard.

To solve this, identify the experiment (in that url of log.optimizely or using Conversion.com Optimizely extension in Chrome)
Archive it, then unarchive it
If last doesnt work, duplicate it, archive the older one (if your results dont matter anymore or you can let them go)