Hello everyone!
I am looking at tools for load testing. Since there are a lot of them, I had to use chatgpt for help. According to given criteria, the list was reduced to 2 tools : k6 and Artillery. The easiest way for us is to use k6. But I would like to know the opinion of people who use or used Artillery. Why did you choose it for load testing or why did you change this tool on another one?
Thank you.
3 Likes
It’s probably not the answer you were looking for, but you might want to look into JMeter as well. IMO, it’s biggest advantage is that it gives you distributed testing capabilities out of the box.
3 Likes
What was the given criteria that reduced the tool choices down to 2?
2 Likes
Hello David!
Our was:
- support for multiple environments (4 or more)
- minimal integration effort in Kubernetes
- open-source (for now)
- community support
Since we use Grafana, k6 was expected as one of the tools.
1 Like
Hello Abhijeet!
Yeah, first idea was to use JMeter 
, but we decided to see if there are other tools which can be integrated in Kubernetes.
Now we have on our plate k6, Artillery and JMeter. Since Artillery is something new for me, I wanted to know opinion of our community.
But at this point, I begin to think I should remove it from my list and consider only k6 and JMeter 
2 Likes
I think you missed a fundamental requirement, whether you have skills to use the tool in the team. Having Artillery as ‘something new for me’ makes no sense from a delivery point of view, unless there is a specific benefit that outweighs the upskill time you’ll need.
I have been frequently forced on to new tools, and even with an extensive knowledge of a range of tools it still took an extra couple of weeks to pick up enough knowledge to make a half-decent stab at a performance test.
JMeter in Kubernetes is pretty straightforward, you just need to consider what you’re using for load generation.
Agree with you. QAs, who will be involved in load testing, do not know neither JMeter, nor k6 or any other tool. So we shoul start from very beginning. We already have different opinions which tool is better for us in our company.
For myself, I didn’t decide yet, even I have used JMeter and only touched k6…
What I will say then, and please take this with the absolute greatest respect, I would recommend you hire a performance tester. Being able to use a performance test tool is maybe 10% of a performance testers role, the other 90% is interpreting what the results actually mean.
Unless of course one or more of you are looking to upskill into the area, then absolutely go for JMeter as the best of the open source tools. You would actually be better off going for Neoload or Loadrunner but they’re commercial products with five figure costs associated with them.
3 Likes
Actually, we had such idea too , but we saw that exiting QAs also would like to grow and to upskill themselves as you said. So we decided to give them such opportunity.
P. S I think this is great that they want to learn something new. 
Yeah, commercial products are not relevant for us now, but I will remember your advice.
2 Likes
There are tons of performance testing tools out there, and most cover the same essentials. Do you have specific tech requirements, like WebSockets, or are you interested in branching out into Chaos Engineering? Both can narrow down your options a lot.
Lately, I’ve been diving into Gatling which would be worth checking out since they offer a free online academy, which could be helpful for your case: Courses
Also, ISTQB has a solid performance testing syllabus if you want some beginner information: https://www.istqb.org/certifications/performance-tester
Another angle to consider is observability. Your monitoring setup might impact the tool you choose, as some integrate more easily with popular observability stacks.
1 Like
Well, long story shortly. At the moment, there is no specific tech requirements and they will not appear in close future. 
Gatling…interesting…I will take a look.
From observability perspective, k6 is better then JMeter. But I have time for search, for discussion and understanding if this or that tool is suitable for us before it will be decided what to use.
@veronica_s please look on your digital footprints of skills which you gathered after investing time and smart work then choose a most relivent and near by tool to rock the solution with stress or load.
Keep in mind that trying a new tool or skill implementation may add latency to your deliverables timelines.
This is one of the reasons I started this research so early. I will
1 Like
Dear community!
Thank you to everyone here for the help, advice, and explanations you have shared.
1 Like
I like both of your proposed tools and have used both in a professional setting. I would suggest having narrowed your decision to a small number of tools as you have, to now start thinking about what you want to get from the tools.
- What are the most important features you want and need?
- What challenges to you expect to face when using them?
From there I would undertake a POC to answer those questions and others you have. Once you have done that picking the right tool should be easy, either they both do what you need and so you pick the one people preferred working with, or, only one of them does what you need and so you pick that one.
1 Like
Keep in mind the ecosystem around the tools as well. Sometimes it’s not just about the core tools themselves but what other tooling or extensions/add-ons from thirty party / community do you find around the tools that are useful.
And how extendable the tools are for your needs should you need customizations for your tests. And what kind of support you can get from the community or the vendor for the tools.
1 Like