In addition, I also set Page level cache option on in Joomla (which the site uses) with 240 minutes TTL.
I was hoping that all of these changes would speed up my website, but the changes are far from significant. When Cache level is set to - Cache Everything in Cloudflare, the performance (speed) of the webpage should mainly depend on Cloudflare.
Most of the users that visit my website are from India. Using GTmetrix.com, these are the results I get when the home page of my site is analyzed:
Using Cloudflare is just not a good idea for my website due to slow response received by visitors in India.
Even if I consider the test results using a server from Texas, for a page that is less than 300KB, it shouldn’t take so long for a page to load
Javascript on my page is not a concern. It loads asynchronously. I tested the page speed by turning it off and noticed that it is adding less than 800 ms to total load time. But regardless of the time it is adding to the total load time, it’s the main content of the website that matters.
There is an opportunity to optimize images and I’m working on it.
My main concern is TTFB and LCP, which increases significantly when my website is accessed from India. The tests done at GTmetrix.com use desktop, but from mobile, it’s even slower and adds about 2 to 3 seconds more.
Will you please provide your opinions on this issue? I am a novice on these matters and just learnt as much as I could in the last 3 days when I suddenly found out about this problem.
I was using www.airportchandigarh.com and later removed www. I don’t know how much of a difference that would make, but after running more tests, I see test reports on Gtmetrix from Mumbai still showing poor scores.
Can you run the test multiple times? Because usually during the first test, the cache is not populated yet. You have to run the test second time to get better result.
Yes, I realize that. I ran it a couple of times after removing www. Since then, several visitors would have already visited the home page. So by now, Cloudflare must have the cached contents on multiple servers.
The recent test is showing some improvement though - which is surprising. Yesterday, I ran many tests and saw no improvement. It’s happening after removing www.
Very possible your old page rule wasn’t covering the www. so this makes sense. But that’s good there’s an improvement, GTmetrix will say how you can improve it further if you wanted
For future reference, if you wanted to cover subdomains and the main domain you would have your rule as *example.com/*. The * at the start will cover both. If you wanted only subdomains then *.example.com/* is your friend.
Thanks and I have made the changes as you suggested. I’m trying to dig down deeper. The results with GTmetrix are not consistent. I have tried Pingdom also but unfortunately they don’t offer a test from an Indian server. There is an eerie wait time of about 1 second before the webpage loads. This is based on personal test done using Safari on my iPhone. Since I’m in India also, it’s easier to understand practically what a typical visitor experiences. Another factor that goes in the speed favor is browser cache which is not taken into account by the test sites, but it can be noticed personally. As I’m loading a page twice, it gets faster. I will increase the browser cache time also in Cloudflare after working out the optimum time based on the changes made to the site.
