Best tool for measuring loading time of pages?

Hi,

In the company where I work we are encountering performance issues when loading pages.

I’d like to make an inventory of loading delays for every page. What tool would you recommend ?

Maybe load testing tools like Webload, Loadrunner etc ?

3 Likes

I’m mostly used to using JMeter (OpenSource, free & easy to use)

What kind of webpage are we talking about? Do you use REST API’s or still SOAP or something else?
Make sure to keep your performance script ‘realistic’ to ensure correct response-time results.

If you want to go nuts use for example JMeter + TICK Stack: InfluxDB 1.X: Open Source Time Series Platform | InfluxData

  • Telegraph
  • InfluxDB
  • Chronograph (or Grafana)
  • Kapacitor

Insane dashboarding & live followup of your results, highly recommended.

3 Likes

I red that Jmeter is hard to set up, it could take time, is it true ?

I’d like to test it directly in the GUI and not just API calls.

1 Like

I don’t think so, it’s actually very easy to setup. (TICK Stack is a bit harder to setup)
I don’t think you need a lot of technical skills to setup JMeter either.

If you are looking for tutorials: JMeter Beginner Tutorial 1 - How to install Jmeter - YouTube

This guy explains it all very well

This will be quite difficult in any performance testing tool.
You can do call to retrieve all html, css, etc but … reporting will look dirty.

I would advise to start testing your API’s. Because if your UI loads slow you still won’t know if it’s due to the API or UI.


I don’t know how many users you wish to simulate of course but I can imagine if you want to simulate 10.000 users, you won’t want to spin up 10.000 browsers :stuck_out_tongue:

1 Like

Yes I understand that testing in the UI is less relevant, but the need of the company is more like from a user point of view. Basically they just want to know what pages takes more than 4 seconds to be loaded.

I agree with you that using the API is the proper way but actually they want to measure the UI performance too.

1 Like

You can always test it manually and look at the response times in Developer Tools.
What sometimes happens is that you use JMeter to set load on the web app with 100-10.000 users and then you go manually through the UI while there is a bigger load.

But yea it will never be a proper result :confused:

Like Kist mentioned, with JMeter you would have the most control of what you can check.

But, if you want to start of with something easy (and more visual, since JMeter can see a bit overwhelming with all of it’s options) just try the Lighhouse from the dev tools to get a general idea:


Here are a few more web-based services which offer similar insight:

https://developers.google.com/speed/pagespeed/insights/

PS the MoT Club website seem really performant :smiley:

2 Likes

Have a look at this tool. It would provide some insight on how the page is in diferent areas.

2 Likes

Thank you guys for all these answers. I saw different tools like “Fast or Slow” but actually when you have to test a software’s company with real login/pw I think we need different tools like Jmeter, isn’t it ?

Basically we have to be able to run a real scenario where you navigate into pages or you call API, not just ping a public page like where we are right now :smiley:

3 Likes

Subject7 can do that! Let me know if you want to learn more, or check out our website here! Software Testing Automation Tool Demo

2 Likes

In the past I coordinated performance tests with LoadRunner and JMeter.

LoadRunner is an expensive tool, but it has also a lot of features. For a normal load test and stress test this tool can provide a lot of information.

JMeter is a cheap tool for a load test. It has less features than LoadRunner, but it might be suitable for a small test.

The most difficult part of the performance test is to interpret the data. Suppose the load time of a page is 5 seconds. What is the bottleneck?

2 Likes

In my opinion, there is no tool better than Apache Jmeter to use when you need to work on load tests.

Being VP Delivery and Operations, at Software Testing Company here is the blog that may help you better understand JMeter Tool: Introduction to JMeter & How to View JMeter Test Results in Real-time?

2 Likes

Depends what you are looking to do?

To monitor production systems with synthetic checks, pingdom, speedcurve, wepagetest, or gmetrics are not a bad idea.

For performance testing prior to release I’ve been moving teams away to jmeter to k6. See my testbash talk next month as to why.

You might also consider an APM solution (newrlic) or analytics (google analytics, amplitude) for real user metrics .

2 Likes

If you’re experiencing universal delays, I’d say that focusing on performance not under load might be your best bet to start with. Mainly because it’s the low-hanging fruit, and you don’t have to get into complicated discussions about acceptable load and delays and time of day and yadda yadda.

I really like Lighthouse’s approach; knowing what components of page loading are slowest helps figure out what to address. It’s also possible to run it from CI.

Also, and I’m biased because I work there, Sauce Labs’ front end performance testing is pretty great :sweat_smile:. It gives you similar response breakdowns as Lighthouse, and breaks things down: What calls are taking the longest, is page performance based on rendering or asset fetching or server waits etc. It also automatically establishes a baseline of performance and tells you over time if your site drops below the baseline, for commit-to-commit monitoring.

If you already have existing Selenium scripts, gathering data just requires you to run them via Sauce Labs and add a couple of desired capabilities. You can then use the REST API to download the results.

I agree with @crunchy that adding APM or analytics might be helpful as well; it sounds like this is causing a business impact, so being able to quantify that will help decide what kind of spend you want to do on the process.

3 Likes

What kind of pages?
What sort of application are the pages part of?
What does ‘encountering performance issues’ mean? page unreachable, page load time out, page stutters at display, page loads partially, page loads fine but in a time that’s longer than desired, users encounter display issues when browsing it?
Is your ‘page’ within the ‘app’ distributed across multiple device types or configurations? Would you have to narrow down some specific states?
Is the performance issue constant or randomly showing up? Does it show when some specific content is called/retrieved/processes/displayed?
Is it server-side hardware, network firewall, some slow external provider, overload of the internal system…

Without knowing anything about your problem, application, and its potential location, it’s hard to recommend a tool.

1 Like

Hi @ipstefan,

I meant measuring the load time of every page of our software, simple URL pages with PHP code behind.
Current issues are more related to the fact that some pages take too much time and some users are frustrated.
Consequently I would like to know what url take more than X seconds to be loaded, then we will investigate.

Load test tools maybe are too much complicated for this purpose and simple tools like Pingdom can’t login into the app and navigate into pages, so maybe something in the middle could fit.
I’m going to try a tool proposed by Datadog. They don’t provide load testing but it could be enough in order to get simple metrics on load time.

1 Like