It seems that you have disabled the cache ?
I understand that you’re using https://developers.google.com/speed/pagespeed/insights/, right?
Did you actually read the sugestions there?
It says precisely:
Optimize images, enable compression, optimize CSS, move scripts to the bottom.
How do you actually expect us to help, while ignoring our tips as well as output op the tool you are using?
Many good suggestions here.
Grav provides a list of common ways to improve performance here: https://learn.getgrav.org/advanced/performance-and-caching
The most important one is a fast hosting with SSD drives. Some images seem to take a lot to load, without going through Grav, but directly loaded from the filesystem. I run a report on the webserver and the response is not optimal https://www.bitcatcha.com/my-report/?r=MTIxMjQ5MzA%3D (getgrav.org got a “B”)
Enable GZip compression, as PSI says.
Enable CSS pipeline and minification.
Optimize the images.
Enable the images cache.
Consider using a CDN for assets https://getgrav.org/blog/turbo-charge-grav-with-a-cdn
thank you for jumping in and for your feedback, we are already using a CDN cloudflare to be specific as mentioned above we have temporarely disabled as needed to fix some mautic related issues, however with CDN enabled speed test didnt seem to show much improvement.
About server speed we are aware as well, we are currently evaluating solutions, we are looking for a hosting provider that support at least php7, SSH, possibly HTTP/2 at the moment choice seems extremely difficult as some of the solutions offer those features but at the same time we are very unsure to the extent of the company’s reliability.
We have been suffering two downs in a month and current hosting explanation was that there have been many requests, we really are not sure about this instance, we are currently accessing at different times (time zones) and only two of us handle the site, so it is very unlikely. (Traffic is basically non existing atm as site’s new grav version is just been launched)
Needless to say we need a solution and fast, we are looking at these netsons, shellrent, siteground do you happen to have experience with any these companies?
Can you perhaps suggest a reliable hosting that can guarantee those features mentioned above and possibly uptime but won’t cost an arm and a leg? Current budget is around $ 200/year.
I think the hosting really depends on your target geographical zone. If it’s Finland, you’d probably better pick a local one. If worldwide, we have a good non-exhaustive list of hosting providers on https://learn.getgrav.org/webservers-hosting.
I’d stay away from shared hosting and use a VPS using https://serverpilot.io, so it’s easy to manage and you can control the power you get.
I’ve had good success with DigitalOcean and Vultr, along with AWS and Google Cloud Platform.
This blog post covers AWS setup with serverpilot.
Also, I would definitely use a CDN. Cloudflare Free plan offers CDN with 116 (and growing) locations, IPV6, HTTP/2 & SPDY, WebSockets, free SSL and more.
thank you again for the insights, as a non-tech guy I must say I thought that choosing grav would simplify things a little bit, so far it seems rather demanding in terms of hosting features.
Still referring to the facts mentioned above, even not been so technical by sheer logic, previous site ran same content with Joomla and on same hosting and was 50% faster on average (without CDN).
As for community it seems really good and young so this might be one of the reasons we will keep it up.
Then again it might be that the current hosting which is based in FInland is painfully slow as your test seem to suggests.
From this experience we will try to steer clear of Finland, (we dont know any other hosting that offers decent compromise of features/pricing) we are opening to new markets anyway, so anything goes as long as can match those current needs. Just wondering are you affiliated with those hosting companies you suggested? Do they offer specific grav support?
thank you for the feedback, as I mentioned already twice we already are using cloudflare, the AWS set up sounds a little tricky to me, I will try to see when dev comes back whether he can shed some light. It looks like a good solution under $10/month if thats real could be taken into consideration, I cant see any features list to compare to other hosting suggested here, I will try to dig more into it later.
Yes, there is a bit of a learning-curve with both AWS and Google Cloud Platform. They both offer free trials though for 365 days, and have good tutorials on learning the ropes.
- AWS has EU locations: Frankfurt, Ireland and London.
- GCP has EU locations: Belgium and London. (They say Finland coming soon!)
- DigitalOcean has EU locations: Amsterdam, Frankfurt and London.
- Vultr has EU locations: Amsterdam, Frankfurt, London and Paris.
I’m not affiliated with any of these entities, but I do have experience running Grav on all of them.
Here is a very thorough article comparing some VPS hosting for 2017 (it’s in English):
If you want to make sure that loading time is not afected by server-side page generation, install
Blackhole module, @tancredi it will generate all sites as plain HTML.
Also, you can download all the graphics and other assets optimized to minify browser load from PageSpeed results site.
getgrav.org is hosted on Linode and we’ve been very pleased with performance.
if you use our referral URL, you’ll be supporting Grav too! https://www.linode.com/?r=300c424631b602daaa0ecef22912c1c26c81e3af
@rhuk thank you for the suggestion, dev in on holidays and I was wondering why the retina ready goes away every now and then? He fixed it 10 times already and now is back to blurry images very frustrating indeed especially cause I cant figure why.
Here is a code snippet for one of the images:
<div class="column is-2-tablet is-half-mobile" markdown="1">!(jou-pet.png?resize=140,109&quality=75)</div>
Do you mind taking a look and let me know whats wrong what should be changed in order to get retina ready images?
I will get back to you on hosting and speed issues when dev comes back.
There is something to be said for using a VPS, your site will scale a lot better in most cases. Though decent shared hosts can still deliver the goods, depending on who you use (ie., no EIG-affiliate or subordinate). My host delivers <500ms load times from the US to Europe (Norway specifically) with a basic pages with images - fairly heavily optimized on my part - and <1s on a heavier page chock-full of images - also optimized for delivery. Images are not specifically optimized for retina-specific delivery, but because of global CDNs will yield optimal delivery of both smaller and larger resolutions (images on Flickr, other assets cached via CloudFlare).
It sounds like your are experiencing non-persistent caching of images, causing them to appear blurry when they should not. This with the proviso that you are indeed testing the site with browser-cache disabled. This would be an issue with the underlying system, ie. how the host handles PHP and the filesystem. I’ve been through a variety of hosts, and apart from a VPS - which in my experience will require more maintenance from a SysAdmin - I would recommend GeekStorage. I’ve used them myself for three years, and for a shared host the quality they deliver is superb. No downtime in that time, great servers (SSDs and all), variable PHP-versions, and LiteSpeed (which as far as I know still beats Apache and Nginx on speed).
In regards to my note on “no EIG”: EIG is a conglomerate/investment group that heavily invested in buying up web-hosts and reducing costs by cutting quality in servers and support. I’ve been through three hosts that either were or became owned by EIG, and though I resolved almost all technical issues myself they were completely inept in solving solutions on their end. Downtime was a frequent issue, and their pricing is pretty much on-line with every other shared host available - its just not worth the time or cash.
@OleVik thank you for your insights and contribution, it was very informative, I fail to understand why on one page they look blurry and another one they still got the retina ready feature, same images different pages on the same website, do you think this relates to the issue you pointed out how the host handled php and filesystem?
Concerning hosting I was trying to evaluate this https://www.linode.com/pricing as suggested by @rhuk but they have so many plans, they claim “non calculator required” but I honestly I’m not sure which plan will suit best for our set up.
We are trying to get some visitors to the website so perhaps some plan that doesn’t take down the site when it starts to have some visits would be nice. Our goal would be about 200k visits per month but ofc it takes time to achieve. Will this linode 2GB be enough?
Are you sure that you call the image the same way on the two pages ?
Because as far as I know:
<div class="column is-2-tablet is-half-mobile" markdown="1">!(jou-pet.png?resize=140,109&quality=75)</div>
Will resize your image, but won’t do anything about retina.
@AmauryH thank you for your message, actually in the other page I see this:
<div class="column is-2-tablet is-half-mobile dd-client-img" markdown="1">!(sonokogroup.png)</div>
I can tell they are different but Im not sure what classes stands for. This one doesnt have that resize attribute though. Can you tell what shall I change or edit?
Ok it seem that the resize and quality attributes cased the issue, if I remove these (resize=140,109&quality=75) I get retina images, Im wondering whether dev used it to optimize images though by taking it out will slow down the site.
I think that you and your dev should read the documentation.
Your images were huge, that’s why your dev resized them.
But if you want to generate one picture for mobile, one for desktop and one for retina (and to me, it seems that it is what you want), then have a look here.
You’d have to take in quite a lot of traffic to actually break or exceed what the host offers, but a service like Linode, DigitalOcean, Amazon, or Google will allow you to more easily scale that for a growing site.
A site with some hundreds to a few thousands of daily visitors should be able to run well on a shared host, but that 2GB Linode-offering will work very well. From both startup to growth it will handle what you throw at it. The whole “no calculator required” is really just indirectly saying that they handle scaling on their end: There is no specific amount of KB that you can guess beforehand when calculating needs; sites change and grow, but with a decent host you can easily change plans and VPS’ like DigitalOcean allow you to attach on resources as needed.
Do you happen to know anything about “requests”? Current host mentioned that they downed our site due to requests as traffic is non-existing right now. I just want to make sure and avoid another disappointment and more damages thats all.
Hope this host and 2GB plan is as reliable as you mentioned.
Well requests are hits on the site, which does imply there is traffic but in the sense that the site is marginally accessed - ie. something is pinging the site/domain but not loading assets. A competent host should be able to tell you what requests are being performed and from where, as it seemingly does not yield actual traffic for the site but only stress upon the server.
The great thing about the larger VPS hosts such as Linode is that reliability comes from their scale, not just yours. Traffic is not just balanced within the server that hosts your site, but the entire network of servers that they host sites on. Should the site spike in traffic/growth, it is allocated to a more optimal server. For example, some such hosts charge you for actual (and minimal) usage, so an increase in traffic creates a proportional increase in your bill from a set rate. Changing between hosts should be as simple a process as possible, where you just decide on a new host, redirect your domain to their nameservers, and am up and going again as soon as domain propagation finishes.
thank you for following up, apparently those were access to site, I’m not very technical about this issues however by using sheer logic, as we have two users accessing the site maybe 30/50 times over a month, if this caused the site to be downed that host is a straight up fraud in my view.
We are considering Linode, so far I have hard times understanding whether databases are included in that 2GB plan, we need two at least, doesnt seem to mention it anywhere. Also DNS management is apparently an add on, I wonder how much that will cost.
As soon as dev comes back we will review this elements and start the transfer, I can myself move the grav site but for analytics and marketing automation I need some support.