[Hummingbird] Asset Optimisation - reCaptcha fails after some time. Possibility to exclude individ

I am using the assets optimisation feature in Hummingbird, with files hosted on your CDN. All works fine, except that after some time (usually a week) on websites using reCaptcha, the file that loads reCaptcha becomes void, resulting in an error.

I am not 100% where the fault is, but I suppose that Google loads some extra files from some dynamic urls, therefore the CDN files are no longer accurate.

The issue gets fixed every time by clearing Cache.

It would be great if there was an option in Hummingbird to remove individual files from being uploaded to the CDN, and still being served from their original source!

Another annoying error, on websites using the Page Cache in combination with the Assets Optimisation, every one week or so, all the files hosted with the CDN fails to load!

https://www.dropbox.com/s/0hq4s96m01irog5/Screenshot%202019-01-21%2009.59.32.png?dl=0

  • Adam Czajczyk
    • Support Gorilla

    Hello Muser Entertainment

    I hope you're fine today and thank you for your question!

    The "503" error screenshot shows what appears to be an error coming from your server, actually. Whether it's a result or a cause of the issues that you're experiencing with Hummingbird's asset optimization - this would have to be checked.

    I would like to ask you some questions then:

    1. you mentioned that this happens "every week or so" - could you please try to "catch that up" and when it happens look into the browser console (CTRL + SHift + I on Chrome and CTRL + Shift + J in Firefox) and see if there are any errors showing up there; if so, could you please take a screenshot of these errors and share with us?

    2. when you "catch that up", would you please tell me exact time and date when you noticed it so we could also check in our logs?

    3. did you notice if those are any specific resources of those optimized? usually in browser console you would see some error with a reference to a specific file, could you take a look, please?

    Let me know about this and then I'd also would like to check this on your server. To do this, I'd need a full access if possible so would you be able to provide me with full access credentials?

    Note: Don't leave your login details in this ticket.
    Instead, you can send us your details using our contact form https://premium.wpmudev.org/contact/#i-have-a-different-question and the template below:

    Subject: "Attn: Adam Czajczyk"
    - Site login URL
    - WordPress admin username
    - WordPress admin password
    - FTP credentials (host/username/password)
    - cPanel credentials (host/username/password)
    - Folder path to site in question
    - Link back to this thread for reference
    - Any other relevant urls/info

    Best regards,
    Adam

  • Muser Entertainment
    • Flash Drive

    Hi,

    I don't think you've read the ticket carefully.

    The assets are hosted on the wpmudev cdn. So how can you say that the 503 errors are coming from my server? Both the screenshots are showing the urls responding with a 503 from the hb.wpmucdn.com server.

    1. I've already attached you the errors showing in the console log: https://www.dropbox.com/s/0hq4s96m01irog5/Screenshot%202019-01-21%2009.59.32.png?dl=0

    2 & 3 All assets, both css and js are responding with a 503. this happened also yesterday at the time of posting the support ticket.

    And you haven't answered to the most important questions, how can I exclude individual assets from being served from your CDN? It should be an option for that.

  • Adam Czajczyk
    • Support Gorilla

    Hello Muser Entertainment

    First of all, I apologize for the delayed response. We've been a bit overwhelmed with questions recently, hence the delay. I'm sorry about this.

    As for the issue.

    I read your post carefully and I stand by what I said. The first screenshot showing the console error points to the 503 error on the resources hosted on WPMU CDN, that's true. The second screenshot, however, points to the Varnish cache and does seem to come from your server.

    However, I did indeed asked some questions that I didn't have to ask as they were actually responded by the first (console error) screenshot. I'm sorry for this, it was a hard day and I must have skipped that specific screenshot.

    Still though, these two things seem to be connected. Now, I didn't say that it's 100% sure diagnosis but that it "seems it's like that" but it looks like there's heavy server-side caching and that is very often a reason for such and similar issues. I know that you shared the debug.log as well but the reason I asked you for the access credentials was for me to be able to check site configuration, check server - especially against cache - and be able to check debug.log while testing Hummingbird settings as the debugging output might be changing too. That's all in order to help you and find a solution as soon as possible as this is my main goal here.

    That being said, could we try again, please? This way this time:

    1. Please double-check (consult with your host if necessary) if there's any server-side cache on the server, especially a Varnish cache. If yes, please try disabling it temporarily, then clear all the caches on site - in Hummingbird best way to do this would be to go to the "Hummingbird -> Dashboard" page and click on "Clear cache" button there (it's important to use this particular "clear cache" button). Then please keep an eye on the site to see if the issue happens again.

    2. If it does happen (or there's absolutely no cache on server), please provide the credentials that I asked about in my previous post here

    https://premium.wpmudev.org/forums/topic/hummingbird-asset-optimisation-recaptcha-fails-after-some-time-possibility-to-exclude-individ#post-1370725

    and I'll include our Hummingbird developers in the case, asking them to check what is happening and how to fix that for you.

    And you haven't answered to the most important questions, how can I exclude individual assets from being served from your CDN? It should be an option for that.

    There's currently no such option, it's either all assets or no assets stored in WPMU DEV CDN. However, I talked to the plugin's lead developer and he agrees too that it would be a good feature so we already added it to the future features list. I'm not able to give you any estimations on the implementation but it's already planned so thank you for suggesting it, it's definitely gonna be a very handy tool.

    Kind regards,
    Adam

  • Adam Czajczyk
    • Support Gorilla

    Hello Muser Entertainment

    The second screenshot, with the 503 error, is from clicking on one of the "hb.wpmucdn.com" urls from the first screenshot. So again, I don't see how you are still thinking that is somehow coming from our server.

    Thank you for letting me know about this, that changes things. The reason I was thinking what I previously said is actually quite simple: the URL in browser is not shown on that second screenshot and you didn't inform me that this screenshot was taken this way. Believe me or not but you're probably the first ever person (I'm referring to the issue reported to us) in long years that actually did something like this - actually clicked on such URL. Literally nobody ever does that so while that's perfectly logical thing to do I didn't expect it and - excluding this case - a screenshot of such error in browser in 99,999% of cases is taken when trying to load the affected site, thus it refers to that site and it's server.

    Like I said, that of course changes things and thank you for clearing this up.

    Just to make things clear on my end too: please understand that whatever I ask or suggest, I got solid reasons for doing this. I'm not here to argue with you or undermine what you say and if you took it this way, please accept my sincere apology! It is not and never was my intention! We're all here to help and even if something that we say sounds weird or irrelevant (or even plainly stupid) to you that always has a reason and the main goal is to help you as fast and as efficiently as humanly possible. But we are also humans so we can make wrong assumptions sometimes or make mistake. I hope that makes sense to you and if I ever made you think that I'm arguing with you or anything like that, I'm sorry.

    Getting back to the point: I asked for the CDN logs to be checked and it turns out that apparently around that time when you experienced the 503 errors there happen to be some unexpected temporary outages of some endpoints of the CDN. Those indeed resulted in such issue for some sites. As I was informed, that's fixed though so 503 shouldn't be happening now.

    However, that is still leaving the other issue - the captcha related - open as it doesn't seem to be related. I assume that since you didn't address my question about server-side cache, there is just none, correct? I spoke to one of our Hummingbird developers about the case and I'd like to ask you for one more small act of help:

    he needs to be able to check "fresh logs" on our end and that means that he needs to "catch that" right after it happens. That being said, could you please keep an eye on the site and once the issue with reCaptcha happens again could you please notify me here right away? You may of course clear the cache when it happens to make it work again but the point is that I'll then immediately let aforementioned developer know about the incident and he'll check what could be happening on our end.

    I'd appreciate that help a lot!

    Kind regards,
    Adam

  • Muser Entertainment
    • Flash Drive

    Here are some fresh logs, from this morning:

    Failed to load resource: the server responded with a status of 403 (): https://hb.wpmucdn.com/leadhub.dk/83d51cfd-7d00-48f3-be97-22793a4048ac.js
    https://hb.wpmucdn.com/leadhub.dk/7adab615-8566-4acf-9276-1975dbcef3c4.css

    The error I get now by clicking on any of the links is "Access Denied".

    Screenshot: https://www.dropbox.com/s/807epshi6j59ofz/Screenshot%202019-02-11%2009.34.15.png?dl=0

    One of my guesses is that the assets sources are being refreshed at some point, but the page cache fails to update them, and still serving the old resources.

  • Adam Czajczyk
    • Support Gorilla

    Hello Muser Entertainment

    Thank you for getting back to me with this info. I've already passed that to our developers so they're checking it. I can also tell you now that meanwhile they found out about some additional issues related to this and are already working for a fix.

    I don't have an ETA but it's already on the table. There's also a chance that we might have a patch even before the update so please keep an eye on this ticket and in case we did have it, we'll share it with you.

    Kind regards,
    Adam

  • Adam Czajczyk
    • Support Gorilla

    Hello Muser Entertainment

    I just got a message from one of our developers that they got a patch for this. Would you like to give it a try? It would be great as it would also help us confirm if it works as expected.

    I've attached a zip file to this post so to apply the patch please replace

    /wp-content/plugins/wp-hummingbird/core/modules/minify/class-minify-housekeeper.php

    file with the "class-minify-housekeeper.php" file from inside the attached zip file.

    Once it's there, please still keep an eye on the site and see if the issue still happens. Let us know, please, if it helped or not.

    Kind regards,
    Adam

  • Muser Entertainment
    • Flash Drive

    I've noticed that this (still) happens when Automate is in place, and the updates are being performed. It seems that the plugin fails to clear the page cache after the maintenance updates, and still tries to serve old minified files hosted on your CDN, resulting in the 503 error on your server.

  • Adam Czajczyk
    • Support Gorilla

    Hi Muser Entertainment

    Thank you for letting me know!

    I just checked and there'd be indeed the case of Automate not triggering HB cache flushing. Automate itself bypasses cache e.g. when doing Safe Upgrade but you are right - it should be able to clear it also.

    I've just consulted this with our developers and I can confirm they're already looking into it. Apparently there are already necessary API endpoints in Hummingbird itself so it's a matter of The Hub development. The case is "on the table" already.

    Kind regards,
    Adam

Thank NAME, for their help.

Let NAME know exactly why they deserved these points.

Gift a custom amount of points.