I have now had the Linux ultimate shared web hosting package for about 3 months.
I have had continuous issues with it since day one that chat support continues to challenge me on and side step.
The biggest issue being server response time when my hosted sites have been idle for an hour or more. I have tried to explain to support since day one how it feels like their server settings are hibernating my shared piece when it is not in use. Then when the sites have been accessed a few times its magically back to loading in 1-3 seconds. So something is going on there with the way the servers have been setup.
This leads to the first few runs/tests (gtmetrix, pagespeed insights, pingdom etc) on my site, after say an hour or so, to be extremely slow.
Gtmetrix literally times out because it thinks the server takes impossibly long to respond and assumes the server is not responding. Here are some links to some tests I ran with gtmetrix:
Where as page speed insights will show that the server will sometimes take 30 seconds to a minute to respond when it's been idle. Forgot to screenshot some pagespeed insights
Attached are some tests through pingdom on one of my two sites that took 30 ish seconds to load. As you can see this page isn't all that big at 4.4mb and the other site I'm hosting is even smaller.
Both sites barely get any traction due to this whole issue. So it's not like im hitting bandwidth caps. And I can see this is not the issue.
I get calls saying the site just took too long to load and people clicked off.
I have gone through everything with support. They even forced me to delete my sites and load blank sites through installation. Even those presented the same issues.
But whenever support looks at it, "oh there's nothing wrong. we cant see anything from our side", and proceed to ignore what I actually tell them. Ignoring all the facts I give them and evidence of the issue. "Well we can't see it now. So it must be your problem. HAVE YOU CHECKED YOUR PLUGINS" or "your pictures are too big and taking a long time to load". Ah yes 4.4mb total takes 30 seconds to load on a site with very low visitors.
Ignoring that I sent them test data of blank freshly installed sites as well.
If I hadn't bought an entire year of hosting up front I would have dropped godaddy and their support two months ago.
Does anyone know who on earth I can speak to to actually get this fixed or get a refund? Since this has been a day one issue that support is ignoring and won't help me or refund me.
Hello @SaltSupreme and welcome to the community. I'm a GoDaddy customer like yourself. I can totally relate to your issue, I hate when sites come up slow.
It sounds like you are on a shared hosting plan, which means you are sharing the same resources as everyone else on that server. It could be a lot of factors when on a 'shared' hosting plan, cause someone's else account could be hogging up on the 'shared' resources on the server, possibly your site is bad, possibly cause you are hosting multiple sites since I know you stated 'sites' (assuming more than one site).
I tried going to the gtmetrix links but had no joy. See screenshot here, https://prnt.sc/x61cc5
I'm curious what are the URLs to the site(s) that you are having issue(s)? Just curious to see if I experience the same issue! Is there a particular time during the day that your customers are complaining?
I totally relate. Honestly, If I'm only doing one website to be up for the occasional person to see/user, I leave my site in 'shared' hosting. Anything more than that and I expect high expectations and my online customers expect high expectations I spin up 'Bussiness' hosting or even 'VPS' hosting, just cause I know I get 'dedicated' resources to those particular environments/setups instead of having to 'share' resources with everyone.
But those are just my thoughts and thinking out loud hopefully this information was somewhat helpful.
Hi there @davisdre317 and thank you for the message.
Yes I probably should have clarified which hosting plan exactly it is https://za.godaddy.com/hosting/web-hosting ultimate package with 2x cpu etc etc
I assumed that if one website on the basic package is okay enough. 2-4 would be okay on the higher ultimate package.
Currently I only have two websites hosted on it chaosandfluff.co.za and behavioursolutions.co.za
Basically they are family websites. I did have my own tech website on there. But I decided it's a serious problem on this package and moved it to my hetzner hosting for the time being. That is also shared hosting on one of their most basic packages. Something like 3/4 euro a month. It does really well there and so do these sites. But it doesn't have cpanel and I do prefer to use cpanel. I used that since I have my animal rescue npo site there and that site on the same package gets around 500 to 1000 views per day without a stutter. Great loadtimes etc.
These sites aren't really going to get a ton of traffic at least for now. But I would like to at least make sure when they do get a visitor, that the person can load the site in a half decent time frame.
Sorry about the gtmetrix data. It seams they wiped it. Basically the only info there was a "read error". Which in this case just means the server response time was so bad gtmetrix decided to stop the test.
Pingdom and page insights will usually run the tests. Occassionally pagespeed insights will give this error:
Lighthouse returned error: FAILED_DOCUMENT_REQUEST. Lighthouse was unable to reliably load the page you requested. Make sure you are testing the correct URL and that the server is properly responding to all requests. (Details: net::ERR_TIMED_OUT)
In other words the server is failing to respond in time
Eventually after a few tests it goes back to working fine.
Essentially what I have done to "fix" the issue is I wrote a script with a timer.
So every 30 minutes it opens a browser and loads both sites and then twice more after 20 seconds have passed. Then it sleeps for another 30 minutes.
This has corrected the issue.
So it really must be something that hibernates unused shared server shards when unused for some time.
Which is just plain unacceptable from their side.
They even suggested at one point my problem will be fixed if i pay them to move my server from Europe to the USA.
Without any explanation why they thought that.
Unless they know their Europe servers are garbage. But just my hypothesis.
Anyway I really appreciate the reply to my thread here.
Hopefully one day they will either actually fix the root of the issue or I will just wait till my year of hosting I paid upfront is over and then move over to a better hosting company like bluehost or something.
Hi @SaltSupreme. Thanks for your post. Regarding the testing you did with support, you mentioned this is also happening on blank installs of WordPress. Did you have any plugins in these installs (some might consider "blank" in terms of content)? If so, can you share what plugins you're using? If not, do you still have the blank sites up or possibly set one up just for testing on a subdomain?
Hi there thank you for the response.
I probably should have said fresh install. So support made me backup and then delete my sites completely and reinstall just the default Installatron version of Wordpress. So as blank as it can be really. I was told by support on several occasions that they can't validate that its not a plugin problem unless I do this (not that they said its still "must" be a plugin problem after I did this or other support will say "sorry cant see a problem from my side right now" without listening to what I am actually saying).
Regardless of fresh install, long server response time were still a huge issue. Even on one of these full delete wipes and re-installs support said it might still be a plugin with no evidence of such.
There is an option to include Jetpack as part of the install, which I avoided.
So the total list of installed plugins with the fresh install was:
Limit Login Attempts Reloaded
Of which by default only Limit logins is active.
Even on my behaviour solutions site it is only these above plugins plus:
Maintenance by Web factory
WP Fastest cache (for testing against autoptimize to see which one does a better job of combining CSS etc)
On a separate side note to add to this. From what i could see on the cpanel side there was no high usage of resources, read write IO errors or database errors when a connection was attempted during these "hibernating" times.
Thanks for sharing that information @SaltSupreme. My guess at what could be happening is that the initial load of your page is slowed by the caching plugin creating a cached version of your website. Then, when you load it again, the cached version is loaded and everything seems fine. I would double check your cache expiration time limit in WP Fastest Cache and see if it lines up with what you're experiencing.
Hey, @SaltSupreme Here are some of the things I noticed/looked at from what I can see. I'm here in the US.
Hopefully some of that information was helpful and/or leads into a better direction. I like @JesseW recommendations for you in his post as well (hope those help you out).
But for the regional data center that you are in and me hitting it here from the US, I thought was pretty responsive/fast for my experience/expectations for a website when loading, but maybe my expectations are lower compare to everyone else?
Yes the sites right now are loading pretty quickly as they are being viewed quite a bit by yourself and others. So the issue isn't popping up right now.
It's only a problem when the sites haven't been loaded for over an hour typically.
So it becomes hard to show without screenshots the poor speeds when the sites haven't been used and on first load or two what they are like when they have been idle for an hour plus.
Yes autoptimize is being used on the Chaos and fluff website at the moment. It helps fairly minimally on the normal page load times at least in this sites case because there isn't really all that much going on especially css, js wise on the front page. Just a touch here and there really.
Also only added it about a week or two weeks ago because support was like "no no all your problems will be solved with a caching app. Use Autoptimize".
Which is of course ridiculous. Caching helps with sites that have a lot of render blocking with js and css all over the place. Not so much on a site like this.
Plus a blank fresh install of wordpress via Installatron on the cpanel side is pretty minimal. I think under 1mb including the images etc.
If that produces this server response time issue then anything will really.
With woocommerce it hasn't been setup at all. Just a default install, no items etc. It's not even really accessible to the user and was also only added recently.
This particular page is for my aunt who makes clothing. Wedding dresses, kids clothes etc etc. And she really just wanted a site to show off her designs and market herself. It's only been a recent addition of woocomerce as she had the thought of maybe taking orders through the site.
So I haven't even though if it as having potentially been the problem, seeing as the issue was there months before woocommerce was installed.
I do also work with Woocomerce a lot as I mange a website for an ink, toner and office stationary company with just shy of 20, 000 products in their store and about 50, 000 views of their store a month. Typically (although not always), the default install is pretty fine settings wise as long as you aren't loading extra funky addons to it that change functionality. Usually you need to watch out for ajax based search plugins and make sure that they can't just run a heavy search of every item tens of times a second.
Both sites are on the same server shard (in the same package hosted through the same cpanel). The actual content load times are slightly different due to the chaos and fluff site being more visually focused so the images are of higher quality.
The other site is more text based and has two images on the front page I believe.
And don't worry it has nothing to do with your expectations.
It really just has to do with the particular error that only causes long server response (but not slow load speeds), when the served sites have been idle for an hour plus.
Thank you for the input and suggestions.
Hi there thanks for the quick response.
However as I said the blank initial installation via installatron in cpanel does not have the caching plugin, and still produces the same problem when the server has been idle from external access for over an hour. If there is no caching at any level enabled and caching was the issue, then it should not produce the issue anymore. However without any caching plugin or caching setup at all, it still produces the problem. Therefore cutting out that as a potential cause.
Also I have only recently in the last 2 weeks been trying out some caching plugins, but this has been a problem plaguing me since the first day I purchased the hosting back on the 25th of October
So it is not a plugin problem at all.
As fullstack dev the only thing that I can think of that would cause this is a server configuration on the core server side in order to manage the shared server shards for each persons shared hosting package.
My thought process is as follows:
1. If the entire server shard is cleared of all my files and reset to default settings within cpanel. Nothing I have done on the server setup side could be the issue.
2. If I create a fresh install via installatron with the minimum setup (no jetpack) and the only plugins automatically installed via installatron are akismet (disabled), hello dolly (disabled) and limit login attempts (enabled).
Nothing I could have added theme, plugin or code wise could be the issue.
3. If the problem persists and the above 1 and 2 are true. Then there must be an issue elsewhere.
Conclusion: If the two areas I have access to (cpanel settings and wordpress installation) are both defaulted as setup by godaddy and can't be the cause as surely the capanel managers on your end know what they are doing setup wise.
Then therefore something I do not have access to must be the cause.
What do I not have access to? The root settings for the server and the server itself.
Therefore there must be something wrong with the main server settings or the must be an issue with the server itself.
Does my login seem sound?
Also surely 30+ seconds (sometimes over a minute) to recompile cached versions of a tiny page with a server shared of 2 virtualized cpu cores and a gig of ram, is totally impossible when the site has no activity.
Just another hole in that theory really.
Also to be clear the page itself doesn't take 30 seconds to a minute to load. Pages take at worst 7-10 seconds to load in the worst situation when the site is idle. The bulk of the time is actually the tests waiting for the server to send the first piece of information or to send a response of any kind really.
So really the throughput speed is no issue. It's just the server response time is just extremely bad in this scenario.
Since I have disabled my script and I haven't been making any edits. Try connecting to behavioursolutions.co.za and see how long it takes.
Assuming I haven't had anyone browsing in the last hour or so.
@SaltSupreme Sorry I wasn't able to get to this sooner. I'm not sure if you've enabled your script yet, but I did not have trouble loading that page. If you did enable it, could you possibly set up a test WordPress site (with no content or plugins) using a subdomain? I can then have it reviewed again on my end to see what could potentially be causing the issue.