Tracker, Bug 1142146 comment #0 We need to see if there are caching issues with MDN, might need some digging around in Zeus.

Today I did some locust.io testing against the top 100 CSS properties' "raw" urls on MDN. (E.g., https://developer.mozilla.org/en-US/docs/Web/CSS/background-size?raw=1&macros) I slowly cranked it up to 180-300 requests-per-second from my machine, and I saw 0 corresponding spike in request throughput on the MDN python processes, per New Relic. (Around 300 requests-per-second I think I might have tripped an auto-block requests threshold, because I started getting lots of "Error establishing secure connection" messages.) Can we check the load-balancer logs for 2015-04-20 14:20 -14:35 to see if it saw a spike at that time? (I need to dive in a bit to see if locust.io is using regular HTTP cache negotiation.)

Yeah, we got about 94445 hits to various /en-US/docs/Web/CSS/* URLs from a python-requests/2.5.3 CPython/2.7.6 Darwin/14.3.0 UA.

Can we tell if the requests went thru to Apache/python, or were they all served by the load balancer before hitting the web-heads?

(In reply to Luke Crouch [:groovecoder] from comment #3) > Can we tell if the requests went thru to Apache/python, or were they all > served by the load balancer before hitting the web-heads? We should have checked this yesterday, but I don't see any hits from the logs I have on the webheads from this IP for any of the docs URLs.

As Luke requested, I've been making some requests this morning for you to analyse. Most specifically, at 10:25 PDT, I made the following series of requests: https://developer.mozilla.org/en-US/docs/Web/CSS/font-size?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/font-family?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/font-weight?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/line-height?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/text-shadow?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/color?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/letter-spacing?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/font-size?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/font-family?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/font-weight?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/line-height?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/text-shadow?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/color?raw&macros https://developer.mozilla.org/en-US/docs/Web/CSS/letter-spacing?raw&macros

:wbamberg - do you happen to know the IP address from which you made the requests? (Still working on running this myself)

(In reply to Luke Crouch [:groovecoder] from comment #6) > :wbamberg - do you happen to know the IP address from which you made the > requests? (Still working on running this myself) No. Do you need me to do it again? Are you having any problems running it, that I can help you with?

So I'm not sure what the action for me here is? :) Would be nice to have heads up _before_ you guys run this so I can grab the logs as they're running vs going and searching for stuff later. Let's re-run this at a time when we're around and can grab stuff off the logs for you.

We (Will, Luke, C and myself) met up today over Vidyo and verified that everything is working as expected. Luke and Will hit some URLs and I verified logs on the Zeus node as well as the webheads and C gave us background music on her lovely keyboard! All good here.