TTFB problems on website

Y

YeOldeOke

Senior Member
Joined
May 21, 2012
Messages
843
Reaction score
24
Location
Pretoria
A WP Woocommerce website (sitting waiting for a response whilst merrily grinding your teeth is not an option if you want to attract customers)

Hosted on what some would say a good host in ZA but they tell me 'sorry, can't help'

Response times of website taken from webbrowser:

Connection Start
Stalled
​14.97 ms

DNS Lookup
​2.78 ms

Initial connection
​21.79 ms

SSL
​8.66 ms

Request/Response
​2.20 ms

Waiting (TTFB)
​5.27 s

Content Download
​46.13 ms



Response time of WP dashboard:

Connection Start
Stalled
​24.84 ms

DNS Lookup
​0.68 ms

Initial connection
​19.65 ms

SSL
​7.71 ms

Request/Response
Request sent
​1.27 ms

Waiting (TTFB)
​4.95 s

Content Download
​46.73 ms


Note TTFB is about 5 secs in both cases, which seems to me to indicate it is not the website but the server. Am I wrong?

Host shrugs, says TTFB not really important. Gimme a break,pls. 5 secs ZA to ZA, could go around the globe 5-10 times. It an age in today's terms.


Before I move site, am I wrong? Is it something to do with site configuration?
 
Unless I'm seriously wrong, I need to move.

I need uptime, loading speed, service, in that order. Cost not really the issue, within reasonable limits. A managed VPS or dedicated server won't really help as the load is not big.

Failover is important for uptime. No 'sorry, we are doing scheduled maintenance' either.

I was looking at Netherlands as the best International location for host? Maybe with a CDN if needed.

I would really appreciate some guidance here.
 
:)

Mesees Google Cloud can be used as a CDN at $0.09/GB. But it takes a techie to figure out how to use it?
 
YeOldeOke said:
:)

Mesees Google Cloud can be used as a CDN at $0.09/GB. But it takes a techie to figure out how to use it?
Click to expand...
hey. I am with cloudafrica. ni issues so far with them. not sure if your issues are server related or not.

also I use cloudflare as free dns service.
 
Hows cloudflare running these days. I noticed a while back there were some issues in ZA, not sure if the relevant website servers were going down or cloudflare was hicupping.
 
TTFB is important, IF it affects your page load time overall.. Some processes you off load server side as opposed to browser side, can increase TTFB but, won't always affect your page load times.. If your website configuration is not the cause for TTFB latency, there is almost certainly a server side issue at play here, most likely, overloaded servers.. Network latency is also a factor but, I rule this out since you have said it is ZA to ZA.. Unless there is some real messes up routing, you should not really see much higher latency than about 50 or so ms..

If you are looking to move hosts and have 100% uptime, I highly recommend AWS for this(If you dont mind the 160ms overseas latency).. They provide an amazing array of services that will allow you to have your website available 100% of the time and allow you to scale as traffic increases to avoid downtime related to load.. All this won't even cost a fortune..

High level overview of example architecture in AWS to achieve this.. If you want a more detailed example, let me know..

You will make use of CloudFront (AWS CDN), Elastic Load Balancer and EC2 Instances(servers).. CloudFront caches all your static content.. Your viewers hit Cloudfront.. CloudFront routes the requests to the Elastic load balancer, which routes the requests to one of minimum of 2 instances which run your website.. If one of the instances or servers go down, Elastic load balancer knows about and it stops routing requests to it.. Each instance will be in a different availability zone or AWS datacenter, which further decreases region related network outages etc..

You can enhance this further by automatically scaling the number of servers if cpu load for example reaches x percent.. once load decreases, it will automatically terminate the newly added instances back down to two.. You can also set a rule that will automatically replace an instance that goes unhealthy..

Of course, this presents some challenges, specifically having a way of keeping all the servers data in sync with each other and centralizing your database.. Fortunately the database part is easily handled by using AWS relational database service..

Locally, I dont know of any hosts that are able of doing the same job as AWS can.. They may exist, I just dont know of them..
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X