It is now set to true for a few opt-in wikis. After a couple of weeks seeing how it goes for them, we may make it global or revert the changes entirely.

As I had mentioned on IRC: a change has user-visible impact (special pages outdated by X hours or disabled functionalities), be careful there. If we want to apply a change for all wikis, it'll require further discussion.

Maybe make an RfC?

In T7429#149132, @Universal_Omega wrote:

It is now set to true for a few opt-in wikis. After a couple of weeks seeing how it goes for them, we may make it global or revert the changes entirely.

Could you please add our wiki (https://minecraftjapan.miraheze.org/) to the test wiki so that we can collect information that is more in line with actual operations?
The content of this wiki is intended to be accessed from Japan, but the average response time is not good, even considering that we are accessing the site from overseas.
If this will improve the page display speed, we will be very happy.

In T7429#148959, @Universal_Omega wrote:

Now that 1.36.0 is done we can go ahead with this by starting with opt-in wikis.

I'd like this to be added to memes.miraheze.org, the load times are not the best, thus, my request to be opted in

So there seems to be a visible load difference on the special query pages (a huge improvement to those themselves) which was a goal of this, it also likely does decrease database loads but I'm not sure how to get actually statistics of this.

In T7429#150583, @Universal_Omega wrote:

So there seems to be a visible load difference on the special query pages (a huge improvement to those themselves) which was a goal of this, it also likely does decrease database loads but I'm not sure how to get actually statistics of this.

You may want to use grafana.miraheze.org and Google PageSpeed Insights to check your server usage while measuring Special pages(ex:Special Wikiforum) with Google PageSpeed.

https://minecraftjapan.miraheze.org/ is fine with manipulating wgMiserMode.
How about comparing on and off with that?

In T7429#150773, @Shili wrote:

In T7429#150583, @Universal_Omega wrote:

So there seems to be a visible load difference on the special query pages (a huge improvement to those themselves) which was a goal of this, it also likely does decrease database loads but I'm not sure how to get actually statistics of this.

You may want to use grafana.miraheze.org and Google PageSpeed Insights to check your server usage while measuring Special pages(ex:Special Wikiforum) with Google PageSpeed.

https://minecraftjapan.miraheze.org/ is fine with manipulating wgMiserMode.
How about comparing on and off with that?

Well measuring Special:WantedPages for example on wikis there is a huge benefit for loading on those particular pages.

@Void: what are your thoughts about doing this?

I think we shouldn't deploy this change globally for minimal, if any true performance improvements, which would also have negative impact on wikis' special query pages being updated less often. I don't think the community would fully appreciate a sudden change like this globally due to the impact. It seems to not have the impact I wanted when creating this task. My opinion, revert the opt-in, remove it from all current wikis, and remove the updateSpecialPages cron from puppet so it isn't being unnecessarily ran. This can be probably be reconsidered later if it is necessary but for now I don't think it is.