Publisher does not support the Fluid field type. Please do not contact asking when support will be available.

If you purchased an add-on from, be sure to visit to add the license to your account here on

Ticket: Significant performance strain on EE, possibly from Publisher

Status Resolved
Add-on / Version Publisher 3.1.0
EE Version 5.3.0

Hop Studios

Jan 08, 2020

Hello Boldminded Team,

We have identified a fairly significant performance strain on one of our client sites, and we are suspecting that Publisher is involved in the slow site load times.

We are looking primarily at the home page of the bcinvasives site, which can be found here:

I took the ‘view source’ and pasted that into a new template, to bypass any EE tag processing, and that template can be found here:

This new template was still loading quite slowly, so I created a new HTML file and put it in the root of the FTP folder. This page ran in a fraction of the time.

I turned on error visibility in Debugging & Output, and then was looking at the debugging for the raw html in the template of the home page.

the Load times are as follows:

Memory Usage: 34.1MB of 256M
Database Execution Time: 0.0226
Publisher Session Start: 9.0328
Publisher Core Boot Hook: 0.0247
Publisher Core Template Route Hook: 0.0003
Publisher Template Post Parse Hook: 0.0021
Publisher Total Execution Time: 0.0599
Total Execution Time: 9.4610

What jumped out at me was the ‘Publisher Session Start’ above. This seems like the most significant item on the performance list.

I was advised to reach out to you guys to see if you knew more about that line, and exactly what that measurement means.

Again, there are no tags on this template at all, so why is Publisher running any functions on this page load?

Hopefully you can help me debug this issue and improve performance.

My next thought was to use the staging server to disable publisher and see if that improves page load times, but I thought I would ask you about it first.

I have added an account on the staging site for you. Debugging results are similar to the production server, but a lot of the image paths are broken which will affect performance.
The same template is on staging & production:

I look forward to hearing back from you about this, hopefully we can get to the bottom of it.

Hop Studios


BoldMinded (Brian)

Jan 08, 2020

Hi, Rowan. It’s mentioned in the documentation that Publisher can increase load times due to the functionality it delivers. Do you have Publisher’s internal cache enabled? The session start hook is where it queries phrases and creates global variables. TBH 9 seconds is very long for the session start hook… usually its well under 0.5 seconds.


BoldMinded (Brian)

Jan 09, 2020

Are you using Structure? How many phrases do you have?


BoldMinded (Brian)

Jan 09, 2020

For comparison, my local dev environment has a hand full of phrases and is running Structure, and this is my debugging output on a fairly small page:

Database Execution Time: 0.0165
Publisher Session Start: 0.1570
Publisher Core Boot Hook: 0.0903
Publisher Core Template Route Hook: 0.0131
Publisher Channel Entries Query Result Hook: 0.0029
Publisher Channel Entries Tagdata Hook: 0.0017
Publisher Template Post Parse Hook: 0.0026
Publisher Total Execution Time: 0.2677
Total Execution Time: 1.2451

Hop Studios

Jan 09, 2020

Thanks so much for the detailed replies. Its great to have such quick and thorough support for addons!

So, no, Publisher cache was not enabled on this site.

I enabled it on the production site and noticed nearly a 10 second performance gain. From ~14s to 4.7s load time.

I am relatively new to the your add-on and also to the client site, so this is great to know about the caching settings.

Thanks for your support and for pointing out the cache settings.

You can go ahead and close this ticket.



BoldMinded (Brian)

Jan 09, 2020

Ah, good to hear. Yeah the front-load of grabbing all the phrases, and potentially Structure pages is a bit of heavy lifting, thus the cache.


BoldMinded (Brian)

Jan 09, 2020

Comment has been marked private.

Login to reply