APO - How to tell if HTML is being cached

Background

I am benchmarking the WP APO against my current set up (Kinsta host, my own custom KeyCDN with image optimization)

Without the APO turned on CF looks worse than KeyCDN (based on pagespeed insights, gtmetrix, pringdom, chrome devtools). The CF WebP conversion isn’t as efficient as KeyCDN which increased the payload size. Also pagespeed is still reporting some images as not being next gen format so I am not sure is that is due to the NS changes not having propagated fully or the CF cache not being primed when hit be pagespeed - either way I don’t see that issue with KeyCDN.

So I am interest to see how the performance is with the APO turned on - but I am not sure if the HTML is being cached - if I am understanding the header correctly I think it is NOT:

cf-apo-via: origin,no-cache
cf-cache-status: DYNAMIC
cf-edge-cache: cache,platform=wordpress
cf-ray: 5e6de036fa8202c8-SEA
cf-request-id: 05f88a765a000002c81da4b000000001

That looks like the APO is on but the page is being served from the origin and not cached.

Could we get an overview of the header response codes, and also any ideas why this is…

The site is tctcstaging.com

Hi,

Your question has been answered before regarding the cf-cache-status: DYNAMIC header. In fact, I have just answered the same question today right here: APO cf-cache-status: DYNAMIC

I have visited your website and got cf-cache-status: HIT. The reason why you got DYNAMIC is probably because you are testing with “Disable cache” enabled in your browser when it should be disabled.

I hope this solves your issue.

Cheers.

@yevgen, I think you guys should add a sticky post to the forum about disabling the “Disable cache” option for testing APO. It would probably bring down the ticket count by a few at least :slight_smile:

1 Like

Yes that totally solves the problem - agree about the sticky post - it would be good to add a quick overview of the CF header response codes for APO as well.

1 Like

It would be interesting to see these comparisons with a “fresh” starting point on each provider. I’m guessing you compared an established setup on KeyCDN to a newly enabled CF/APO combo.

I’m not saying Cloudflare is unbeatable or doubting the results. Maybe it would be the same either way, but it seems like an unfair comparison when CF/APO is going in blind while KeyCDN has had time to sort itself out through automatic and manual tweaks.

The webp conversion is interesting too. I’m trying that here with CF/APO and I think it degrades image quality too much, even though the reduced size is nice. Do you think KeyCDN is actually doing a better job (reducing file size without losing as much quality) or is it simply compressing it more?

I agree its definitely not a completely fair comparison - however I did try to give CF every chance which is why I turned on image processing and any other optimization that I thought would help.

I haven’t done anything particular advanced in KeyCDN - its largely based on the emails they have sent out for new features. The two main things it has are the image processing, and origin shield - which is required for image processing as it means that once the WEP is created all of the POPs will pull it from the origin shield servers rather than each one having to create it own WepP.

I have some fairly exacting colleagues when it comes to image quality and no one has commented on the image quality since we started using automatic WebP conversion. I guess ultimately its a personal decision as to whether the image quality is acceptable.

I am going to retest the CF performance this week as I had only just switch the nameservers on Friday over and I wanted to make sure it is fully propagated.

This topic was automatically closed 5 days after the last reply. New replies are no longer allowed.