any easy to use utility/tool/profiler/benchmark that able to test what is the maximum users a web application able to support by analyzing session size , cpu speed, memory size..etc and 'PREDICT' when server is overpacked/overloaded?
Apache JMeter is a simple-to-use system that fires requests to your web server. You can try different numbers of users (i.e. sessions) with different numbers or types of requests. You could try increasing the number of users until the latency/speed of your system is unsatisfactorily slow.
Related
We have a Quarkus Rest service, and the client is using org.apache.http.impl.conn.PoolingHttpClientConnectionManager with following settings
connMgr.setMaxTotal(20);
connMgr.setDefaultMaxPerRoute(6);
How would from service we can check if the service support maximum 20 connections?
By default what is the maximum connection allowed in quarkus?
As far as I know quarkus is not limited you can configure a limit with this property but I would not suggest going over 100 connections never, because then you might run into the memory limits or cpu limits. This is why you replicate or scale your backends, so instead of having one jvm handling 150 connections, you have three smaller jvms handling 50 connections each one so you gain some high availability and fault tolerance.
If you want to test the behaviour in concurrency of your application you can always run a load test with Jmeter or other tools which will allow to simulate the load that you want and you would be able then to check the response time of your backend or if you run into resource bottlenecks or other issues.
This is for e-commerce project where the number of users login will be more. I have been given a benchmark 8000 concurrent users need to login and response time should be 3 minutes
#abi , hi .
Let me provide couple notes here.
Depending upon Your connection bandwidth , from my experience as performance test engineer, I'd say jmeter single instance usually holds up to 1k(1000)- 2k(2000) in best case users load.
Considering You have a requirement for 8k (8000 users) load, You need to launch jmeter in distributed mode ( master <-> slaves).
For this config setup I'd recommend to go with 1 master node and 4slaves. For that - You will need 5 machines (aws/azure, whatever) in the same sub-network.
Re more technical details on distributed setup, please take a look:
in public jmeter documentation
please also look into this step-by-step setup manual
Also, when i've been doing set-up for 10k load for one of my recent projects - I did couple notes for myself in g-doc . Let me know if it opens fine for You.
Last note, If You need to do some load/performance tests on APIs that require AUTHZ, I'd recommend to split authorize (IDP bypass) and performance scenario itself - in different thread groups. As usually IDP in DEVs/Stagings does not hold much load .
So at first You need to authorize w/o any load (1st Thread group).
And in 2nd Thread group - start calling target APIs under the test.
It depends on:
Your machine specifications (CPU, RAM, NIC card, hard drive, etc.)
The nature of your Test Plan (number of requests, size of requests/responses, number of pre/post processors, assertions, timers, etc.)
Response time of your application
So if your test is a simple GET request which returns small text response - you might simulate 10 000 of users on a mid-range modern laptop. And if your test is connected with heavy requests, large responses, file uploads, etc. - it might be 1000 users.
Make sure to follow recommendations from JMeter Best Practices
Make sure to have monitoring of resources usage of your system (CPU, RAM, Swap, etc.). You can use JMeter PerfMon Plugin for this.
Make sure that your test behaves like a real browser
Start with 1 virtual user and gradually increase the load until you reach 8000 virtual users or JMeter starts lacking resources, whatever comes the first. If you can simulate 8000 users from a single machine - you're good to go. If not - you will have to consider Distributed Testing.
I have requirement of test api with 10k users. What I choose is :
Jmeter
Linux server to install Jmeter and throw load with 10k users
One node for now
Will perform API with operations like :
Login
Book hotel with post parameters
Update Booking Details
Save Booking
I am thinking to use above for API testing with 10k users, Above tools are enough or I should look for other options like loadimpact, loader or blazemeter?
If you are talking about API you should be rather considering "requests per second" rather than "users" as I strongly doubt that end users will be sending requests to API endpoints via curl or Postman.
No matter whether your goal is "users" or "requests per second" it is only you who can answer as it depends on many parameters like:
your machine hardware specifications
software specifications (OS/JVM/JMeter version and architecture)
nature of your test (request/response size, number of pre/post processors, assertions, etc)
So you should act as follows:
Make sure you're following JMeter Best Practices
Make sure you monitor baseline health metrics of the node which is running JMeter (CPU, RAM, Swap, Network, Disk), you can use JMeter PerfMon Plugin for that .
Start with minimal load (1 virtual user or 1 request per second) and increase the load until your machine starts swapping or any other health metric exceeds, say, 80% of maximum available capacity. Once it happens take a look into active threads (Active Threads Over Time listener) or throughput (Transactions Per Second) - this is the maximum number of users or hits per second you can produce on particular this host for particular this test. If it is enough - you're good to go, if not - you will have to switch to Distributed testing
See What’s the Max Number of Users You Can Test on JMeter? article for more details.
The whole answer is an elaborate it depends.
The best thing you can do is to run PerfMon agents on the server generating the load as well as on the server running the system under test.
This way you should see (in the CPU utilization and free memory statistics) whether you had maxed what the server providing the API can do or whether it is your load generator running out of steam. In the first case you got some readline based on the hardware and configuration you had run with. In the second one you have an indication to employ more than 1 box to generate the load or to investigate settings and options.
Have a closer look at PerfMon JMeter plugin for exact details.
We have a clustered WebSphere environment and have 4 nodes in the cluster. I am trying to find number of concurrent users that use the application. I have turned ON Performance Monitoring Infrastructure (PMI) in WAS for Servlet, JVM and Thread Pool. When I monitor using Tivoli Performance Viewer, I believe I need to look at "LiveCount" under the "Servlet Session Manager". But the count seem to be VERY high, more than what I was expecting (LiveCount shows as 80-100).
Is this the Metric to look at when trying to find the total number of concurrent users?
Does it keep true login count or is it keeping track of number of sessions?
I was told the underlying application creates only 1 session per login until it gets timeout. At which point, the user will have to login again. So to me concurrent user count = session count = this Servlet Session Live Count.
Can anyone help me here and let me know which metric I should look in PMI to get the concurrent user count? We are are WebSphere 7.x.
I believe I need to count this metric value for all 4 nodes to come up with the total concurrent users.
You may consider logged on users vs concurrent users, depending what you are really looking for.
The logged on users would be successfully authenticated and having valid session in the application, but... they may currently do nothing in your application, or abandoned your application but not logged out.
They will be represented by LiveCount - The total number of sessions that are currently live.
The concurrent users would be users simultaneously access your application. They will be represented by ActiveCount - The total number of sessions that are currently accessed by requests.
And yes, you have to sum this from all your app servers.
See also:
Servlet session counters
We have built a web services to serve map tiles like google map based on asp.net.
And The client require that the response time for 1000 concurrencies requests must be less than 1 seconds.
Now we use the loader balance hardware,We deploy the service to 4 servers using iis , then we use the loader balance hardware to distribute the requests to different server.
However someone suggest that we should not use the loader balance,since the browser request limits.
It is said that for a given domain,the number of the requests the browser can sent at the same time is limited(maybe 10 or more).
So we should make our client application request to different tiles server directly.
Now,I am confused,which is the right way?
It is said that for a given domain,the number of the requests the browser can sent at the same time is limited(maybe 10 or more)
This is only sort of true. Most browsers won't make more than a few request to the same domain. However, there is no set standard or defined limit and it is often configurable.
How do you know all your users will be accessing your service through the browser
What happens if you have 1000 concurrent users?
Use a load balancer