I want to load test a URL by hitting it few hundred times at same millisecond . I tried JMeter but I could hit 2 request at same millisecond. This seems to be problem that my machine cant create threads fast enough . Is there any solution to the issue ?
In JMeter you can use synchronizing timer setting it to 100, this way all threads will wait until there are 100 available and hit the server:
http://jmeter.apache.org/usermanual/component_reference.html#Synchronizing_Timer
Another solution is to increase the number of Threads so that you hit this throughput.
In next coming version (2.8) of JMeter you will be able to create threads on demand (created once needed).
Anyway hitting few hundred times at same millisecond is a high load so you will have to tune JMeter correctly.
Regards
Philippe
JMeter uses blocking HTTP client, in order to hit the server at the exact same time with 100 reqeusts you need 100 threads in JMeter. Even providing that, you still don't have 100 cores to actually run such code at the same time. Even if you had 100 cores, it takes some time to start a thread, so you would have to start them in advance and synchronize on some sort of barrier. And that is not supported in JMeter.
Why do you really want to run your server "at same millisecond"? An ordinary load test just calls the server with as many connections as possible, but not necessarily at the same time. Moreover, sometimes you are even adding random sleep between requests to simulate so-called think time.
As per Philippe's answer, JMeter does in fact support synchronous requests. But maybe for what you want something like Apache Bench using -c100 (or tune it to whatever works) is a better option? It's pretty basic stuff but then the overhead is a lot smaller which might help in this situation.
But I would also steal from Tomasz's answer and echo his concern that perhaps this is not really the best way to approach load testing. If you're trying to replicate real life traffic then do you really need such a high level of concurrency?
You need to use Jmeter-server and a host of client machines for load generation. Your single machine is not enough to generate the load itself.
Related
I'm trying to test my https endpoint with JMeter. I want to make at least 10000 requests per second, but when I set the number of threads to 10000 I get way less rps, around 500.
I've tried setting the number of threads to 1000 and 100, surprisingly I get this same number of rps. I'm using HTTP Sampler and "use Keep-Alive" is set to true. When I look in the statistics I see that when using 100 threads, it makes use of Keep-Alive and connect_time is around 100 ms, but when the number of threads is higher connect_time grows, it's like it stops reusing the connections.
I know this isn't a server issue, because I've tried testing that same endpoint with Yandex.Tank and phantom and it can easily maintain 10 000 requests per second, the problem is it can't use response data to make furhter requests, that's why I have to use JMeter for this task.
This can be done by using "Stepping thread group". It will allow you to send 10000 request per second upto specified time. Refer below image.
Stepping Thread Group
Download jar from below link.
https://jmeter-plugins.org/wiki/SteppingThreadGroup/
I hope you are trying to achieve this using one machine. Try with multiple machine or jmeter distributed mode.
https://jmeter.apache.org/usermanual/jmeter_distributed_testing_step_by_step.pdf
https://www.blazemeter.com/blog/how-to-perform-distributed-testing-in-jmeter/
https://blazemeter.com/blog/3-common-issues-when-running-jmeter-scripts-and-how-solve-them/
I am assuming that it is the issue with machine which is not able to generate that much load. Usually, i have use max 300 threads per machine but it depend on the machine config. Just check if the machine is having issue and multiple machine is able to generate more load, considering server is not having any issue.
Hope this helps.
Update:-Usually 200-500 can be handled my modern machines.
Please check the below link to have some more info:-
1.How do threads and number of iterations impact test and what is JMeter’s max. thread limit
2.https://www.blazemeter.com/blog/what%e2%80%99s-the-max-number-of-users-you-can-test-on-jmeter/ .
I'm load testing a system with 500 virtual users. I've kept the "Ramp-Up period (in seconds)" option to zero. So, what I understand, JMeter will hit the system with 500 virtual users all at the same time. Please correct me if I'm wrong here.
Now, the summary report shows the average response time for the first page is ~100 seconds!. Which is more than a minute and a half of wait time. But while the JMeter is running, I manually went to the same page/url using a browser and didn't have to wait for that long. It was not even close, the page response was almost immediate for me.
My question is: is there any known issue for the average response time of the first page? Is it JMeter which is taking long to trigger that many users?
Thanks in advance.
--Ishtiaque
There is no issue in Jmeter related to first page response time.
Summary Report shows all response time details in Milliseconds, the value "100" seconds have you converted milliseconds to seconds?
Also in order to make sure that 500 users hit concurrently, use Synchronizing Timer.
Hope this will help.
While the response times will be accurate, you need to consider the affect of starting so many threads at once on both your server and your client.
500 threads to start at once is not insignificant n the client. If your server has the connections, it will start 500 threads as well.
Ramping over a period of time is more realistic loadwise, but still not really indicative of server capability until the threads have all started and settled in.
Databases can also require a settling in period which can affect response times.
Alternative to ramping is introducing a random wait at the start of each thread before firing the first sample. You can then choose not to ramp over time, but still expect resources on the client to suddenly come under load and change the settings if you hit limits. This will make the entire run much more realistic of typical behaviour. However, you need to determine if your use cases are typical.
Although the heap size is increased, i notice there is still longer time as compared to actual response time. Later i realised it was the probe effect (the extra time a tool generates due to test execution)
I am testing asp.net website using Jmeter. I have used below scenarios to load test. Scenario 1 give me correct result(What I expect and can be wrong) and Scenario 2 is not giving same result. But I have used same number of requests within same time. Can someone explain me why is this?
Scenario 1.
Scenario 2.
Ramp up time does not determine when any of your tests are going to complete. It only controls when your test is going to start.
Also, the number of threads any test can create concurrently is limited to the memory you've allocated to JMeter. Even though you've set the thread count to 60000, if you've hit the maximum memory you've allocated, the threads will either queue up or never generate (you can watch the JMeter logs for thread creating or errors).
I recommend tuning your JMeter instance so you have some stability to your tests, here's a good guide. LINK
No of requests you have sent might be same. But the concurrent user load on the server is completely different.
I had clarified similar question few weeks ago. You can check the answer here.
Check Here
I had used JMeter for load testing but was stuck at a point where had to execute parallel requests which does not seem possible with it.(if possible kindly explain)
So is there any open source tool which could help. Have to increase load upto at least 1000VU
Depending you your hardware you may need to consider remote (distributed) testing where one master host orchestrates multiple JMeter slaves to create immense load.
In regards to "concurrency" JMeter offers 2 test elements to control load pattern.
Synchronizing Timer - for firing X requests at the same moment of time
Constant Throughput Timer - to set exact load like X requests per second
which does not seem possible with it.(if possible kindly explain)
Increase the number of threads(users), that would be the number of simultaneous requests. The home page shows the exact example:
http://jmeter.apache.org/usermanual/build-web-test-plan.html
Hope this helps.
Feel free to try Yandex.Tank - https://github.com/yandex-load/yandex-tank. It can make to 80k parallel requests per second.
As Mr. Dmitri said, You can consider the distributed testing using jmeter.
Also consider the following points,
1.Use 64 bit JVM and increse the heap memory.
2.Run load testing in non gui mode and remove all listeners (you can add it after testing to display results)
3.Use latest version of Jmeter(2.11), because of optimal resource utilization (CPU,Memory). OR
Use Gatling load testing tool(http://gatling-tool.org/). You can find an comparative study of gatling Vs Jmeter
If suppose I will run load testing with 5000 threads or may be more ,
Will the main server under test crash at a certain level.
As EJP said, it is JMeter purpose to find the limit of tested application and how it will react under performance.
So yes it is perfectly possible.
You should read JMeter Manual.
If you are doing any kind of performance tests (load, stress, soak etc) you will want to know at what point your application server falls over i.e. its breaking point.
Once you've found out what your upper limit is, start dialing back the number of threads until you find your application's "sweet spot" for example CPU usage <80% & >70%