Simulating 2000 simultaneous connections, reliable apache scanner?

Question

I need to download a test or simulate 2K connections to my web application to see how it handles the load, etc.

If I run the apache scanner on my laptop, I assume that it is impossible to simulate?

Does this mean that I have to start a separate virtual machine to get to the production server, each of which can be 200 separate connections?

Will I make 1 request x 2K concurrent users?

Or can I play approximately 500 x 400 to get the same results?

Answer 1

You have to check the Bees with a machine gun This is very customizable and allows you to spin a swarm of micro EC2 instances that "attack" your server. This is an exceptional test in the real world. You can experiment with 2000 simultaneous requests, 200 users with a request of 10 each, 2000 with a request for 1 resource each, etc. You feel great where the bottlenecks are.

As for the cost, for copies it costs a penny. This is a great tool. I know many applications and sites with high traffic that use and rely on it.

Answer 2

It is very easy to get misleading results with ab if you really don't know what you are doing.

If you need to simulate 2000 simultaneous connections, you need to open 2000 connections at the same time - simply. If you need to simulate 2000 simultaneous transactions, you need to open 2000 connections at once, and each of these connections should simultaneously issue a request. One virtual machine in the cloud can easily open 2000 simultaneous connections with Linux or Windows. Regardless of whether your software (e.g. ab) can handle it, it will depend on how efficient it is.

However, if you really ask: “I need to simulate 2000 concurrent users,” then the answer will be different: since these users may or may not open connections when they are not in use, and most connections will be idle most of the time.