[Resolved] Toolset is slow on both front end and backend. I saw this reported by other user

[Author]

Posts

[alexV-8]

Look if you compare this page, with toolset - but not used on this page)

see speed test link
hidden link

To this page
hidden link

You can see without toolset is much faster. I have toolset on about 10 sites and I realized they are all slow compared to non-toolset sites. The sites are hosted same provider/location digitalocean, and the VM specs are better on nursely.com I saw someone else state the same thing.

[Christian Cox]

Hi, adding Toolset and other plugins will always have an effect on the load time of a site without a caching system in place. Some performance drop is expected and required. In comparing these two sites, I'm seeing a difference of around 2 seconds in total page load time.

First, I would like to point out that we're not comparing the same site with and without Toolset here. We're comparing two different sites. So it's not really an apples to apples comparison to say that site A loads two seconds faster than site B because of Toolset - there are too many other variables at play, including different plugins and themes, and different amounts of content in their databases. So comparing the load times here isn't a true representation of the difference added by Toolset.

The performance grades given to both sites are actually very close at 77 and 74. The main differences I see are in the number of requests and the initial time to first byte. The number of requests isn't something Toolset can help with much, because it involves assets required by your theme and other plugins, like JavaScript and CSS files. These numbers are different not only because of Toolset but because the sites have different plugin and theme requirements.

The time to first byte is a difference of about 1.4 seconds, with a wait time of 1.82s vs. a wait time of 0.378s for the initial HTML of both pages. I'm attaching screenshots showing where I'm getting that information. I don't see anything abnormal here at first glance. Again, it's not really a fair comparison because these sites are very different. If you'd like to trim the 1.82 second wait time on nursely, a caching plugin is probably worth investigating. A caching plugin could make the initial page loads faster - but it still won't improve the performance grades much because you're not being penalized for initial page load times. You're being penalized more for lack of asset consolidation and browser caching...and you're being penalized for these same criteria on both sites.

If you'd like, you can install Query Monitor or a similar plugin to monitor the performance of individual queries generated by each component during the request lifecycle. If any query required by a Toolset plugin is consistently running longer than around .5ms, then we can investigate and find out what's going on. Please let me know if that's the case and I will help escalate that issue for further investigation.

[alexV-8]

What a great response Chrisitan. I knew it was not apples to apples but I thought it was close enough. And I saw another user post about this issue so I jumped to conclusions. I just ran the test with toolset installed and got 3.77 seconds and deactivated toolset plugins and got 2.94 seconds running pingdom speed tests a few times so it appears to be a 13% drop in performance on a page that is not even using toolset. I think toolset is amazing so I don't really want to do something different, but as you are probably aware that is a fairly important .8 seconds.

I noticed that for some reason knockout JS seems to load slowly each time even though it is only 21k. Maybe a fix would be to disable it from toolset and use a CDN for example. I know toolset does a lot but it seems like if it was more like a .4 second hit on a page not using it hidden link or close to 0 on a page not using toolset would be ideal, including admin pages. For example, I don't need knjs on my home page. I realize that this is probably not an easy fix, just a discussion pont. It's not critical for me today but we are just hoping it will be a critical problem in the future 🙂 We will play around with caching and any JS / CSS bundling plugins to see if we can get something better. Thanks for your response ~alex

[Christian Cox]

Thanks for the feedback, I understand that performance is a major concern for SEO and page rank. I will pass your recommendation along to the developers regarding knockout.js over a CDN and smarter enqueueing of required files. I hope you have good luck with caching and bundling - I think you'll see a noticeable improvement on page load times with those systems in place.