Clojurians Log v2

@guswill That is actually something I am noticing also.

I haven’t invested time into this yet, but probably soon I will have to. Can you share the template config you have used?

@yogthos Can you think of anything that could cause that?

my guess would be some of the middleware in the default template, one possibility could be format wrapping

the drop @guswill is seeing seems pretty significant, I haven't benchmarked recently but it used to be within the same order of magnitude before

so perhaps one of the middleware libraries that's currently in use is creating a performance hit now

taking a look in wrap-base in the <app>.middleware namespace would be the place to start, and maybe just commenting out different pieces of middleware to see which one is the culprit

Hey, sorry for dropping the bomb like this. On second thought I should've more more careful and shared reproducible benchmark results, but I was really strapped for time.

I will try to find some time this week to get to the bottom of this.

The backstory: we used luminus for an internal tool and, after we deployed to our cluster, a colleague complained about latency. I did a quick benchmark with ab and got something in the low tens. I then spent an afternoon debuggingthinking the deploy had somehow gone wrong.

At the end of the day I decided to try the obvious: benchmark on my local machine. I got pretty much the same results.

Anyways, I'm going to set aside time this week to benchmark different template combinations and report back.