Large Scale Central

Website performance and the tale of the graphs

Historically, LSC has been slow performing at night, starting about 5pm and ending about 930pm (Eastern). This is due to many more people showing up at night, and putting additional strain on the server. This was one of the reasons I was looking for new software.

The new site seems to be performing well, I have noticed no load differences, but I couldn’t actually quantify it. So, over the last two weeks I have been performing an experiment. I have had a service putting far more traffic into LSC than it normally gets. For the first couple of days I kept increasing the load to 300x the traffic. You can see the traffic increase on the left hand graph. The two right hand graphs are the CPU percentage use, and CPU load on the website. They barely notice.

5 Likes

Well that should take care of the slow pewrformance issues that have been mentioned. Be nice if more people were her at times and more posts were being made!

Yep, the two big issues with the previous software (from users’ complaints) were performance/timeout issues, and attaching images. This software seems to have solved both.

Historically, we’re over last year’s traffic anyway, and there has been a slight uptick (compare solid line (this year) to dotted line (last year), roughly corresponding with the site changeover near the end of Oct. Ive chopped off the big spike at the end, so this graph is only through mid-November

image

1 Like

For most of the time period where you were having performance issues I was on an old overloaded PC, so my local issues were worse than your performance issues - thus I never really saw any problem compared to other sites.

About I year ago I brought my PC technology up to speed, but still didn’t really notice much of a problem, perhaps some lag, but I was still used to that from the old PC.

I can say that this new software is lightning fast on the user side. I’ve go pretty fast internet and now pages and photos load as if they were local.

Kudos on your diligence in finding the right package and your dedication to keeping the data more-or-less intact!

3 Likes