I have a site that uses about 40 different node-views variants, each with selection rules that include things like node type, taxonomy, Ubercart price, node author roles, user viewing the node role and country.

Will I start to run into any performance issues soon? I have considered splitting them up into multiple panel pages paths like content/$nid, products/$nid, groups/$nid, etc to skip doing all those calculations to figure out which variant to use. Would this make a noticeable difference (or a difference at all)?

I guess this can be summed up in one questions:

How many variants are too many variants and what can be done when that many variants are actually needed?

Comments

Tiuya’s picture

Tiuya commented

Just an update - I haven't done any testing yet, but I know I tend to always have so many variants so I'm making the pages [content-type]/nid and using Page Manager Redirect to redirect the actual node pages to the panelized version of them, in case someone figures out it's drupal and decides to replace 'product' with 'node' to see what happens.

merlinofchaos’s picture

merlinofchaos commented
Status: Active » Fixed

The performance hit comes in the amount of time needed to test each variant. Luckily the tests are all quick. I haven't actually run any numbers to see, but at 40...yes it is likely there is an impact. The bulk of each variant is not actually loaded until the panel is rendered, though, so I don't think it will be that bad. I don't know how much impact it actually is. It might be interesting to run some timing tests just to see.

Since it's a first-come first-served type of test, the variant that gets picked last will show the biggest performance degradation. Whenever you visit a page that shows the first variant, you will see the least.

Automatically closed -- issue fixed for 2 weeks with no activity.