New Relic with w3tc and high Time to first byte

v3
#1

Reproducible 100 % of the time on a new ee stack with a wpfc.

If I install new relics php agent my ttfb results on webpagetest.org are 2-4 seconds, stop the new relic php daemon and disable the new relic monitoring in w3tc and back down to reasonable rate (.1 to .4).

Another side affect of this setup (ee, new relic, wpfc), xml sitemaps by yoast have malformed xml, reproducible 100% of the time. (turn off new relic and it works fine).

This sounds like a new relic issue and not a ee issue, but i bring it up here because ee stacks with new relic are probably the norm, no one else is experiencing this>?

#2

If your full page cache is active and working correctly, I think new-relic/php performance should not matter.

We don’t use new relic ourself much and I also believe it is not a norm among EE users.

On rtcamp.com, we are using https://rollbar.com/

#3

thanks for the heads up @rahul286

i am communicating with new relic support about this and will try and update this post if they provide insights that might help others.

rollbar.com seems interesting, but new relic seems to offer a lot more features important to me and outside the scope of just error tracking.

#4

I did not mean to compare newrelic and rollbar. They are different tools.

Rollbar gives us good insight into error log which I believe is quite useful.

NewRelic is good but I always found it doing “too much”. May be NewRelic should create something specialized for WordPress!

By the way, for server side debugging, I use https://wordpress.org/plugins/query-monitor/ coupled and also ee debug --php flag sometime.

#5