[Music] foreign [Music] systems are able to handle how long it takes to respond to this request and what percentage of them are failing let's take a look at how you can do load testing for your apis and cost man using the collection Runner um in the movies API collection and I'm going to click the collection run button I'll navigate to Performance and here I can see I have three configuration options the first one is virtual users this indicates the number of users I want to concurrently make a request to my API endpoints I can make this 50 or just give it a strengthen and this can be as high as 500. for the test generation this is the duration I want this virtual users to concurrently make a request to my API endpoints for it sets you as 10 minutes but I'm going to make this one minutes for the sake of this video for the load profile you have two options we have fixed and ramped up fixed indicates that we want number of virtual users that's concurrently sending requests to our API endpoints to be fixed all to the specified situation where ramped up means we want the number of virtual users that is indicated to gradually increase for the ramp up duration of specified below so if the first iteration was five minutes for example I could make the ramp operation two minutes and this would mean that between the first zero to two minutes of the test duration the number of virtual users are going to gradually increase from 0 to 50. I'm going to leave this as one now and I'll leave the test duration as one minute as well I'll run my performance first now I could see that some graphs are starting to populate on my screen and at the top here we have the total request sends which indicates number of requests that has been sent in total we have here the request per second which indicates the throughput we also have the average response time which is constantly increasing as more requests are sent which is constantly being adjusted as more requests are being sent we have the error rate here which indicates the percentage of errors that have occurred now moving to the graph in the horizontal line here we have the timestamp for each of these requests that are being sent on the vertical axis to the left it indicates the error reads where the vertical axis to the right indicates the response time for each of this line we can choose to toggle which ones you want displayed or which ones we want to begin for the virtual users it is indicated in Gray and we can see that it is gradually increasing with time so I could choose to click this um so hide or click it again to display it for our throughput which is measured in requests per second I could also choose to toggle this by clicking it or hide it by clicking it once again I could do the same for average response style and I could do the same for the errors so I'm just going to keep everything on the screen and go over them so the first one is the virtual users we set this as RAM dock to be ramp up in the first for one minute so it moves from 0 to 50 users in the first one minute I will see that for the gray here it starts from zero and it goes all the way up to 50 eventually here it splits 48 at hide this uh for the throughput we can see here that at this point we have 13.5 requests per second and we also have an average response time of 811 milliseconds and the error rate itself we could also check for here that the throughput was 42.5 requests per second and the average response time was 741 milliseconds and the error rate here was 1.176 percent I could also take a look at the average response time for example I can see that there was a spike here and it took 1.18 seconds um to respond and I could see that here it seems to be low it took 555 milliseconds and here it was 648 27. for the errors I could see that there's a spike here and the error which is four percent I could say that here no requests filled the request field yes as well and here we have 648 research we have an error rate of 4.545 and an average response time of 648.27 I could also check the performance details for each of my requests here for the create movie endpoints I have 456 requests in total the throughput of 6.96 I could also check the average response time um the minimum time it takes for and it took for a particular request to get sent the maximum time I could see the 90th percentile and the error percentage I could do this for other requests as well at the top here I have options to filter this graph by request so I have to filter by any of these requests to just see a graph for that request itself I could also filter by average response and check the minimum response maximum response and to different percentiles like if I want to run this again I could click this run again button I complete resets to reset the filter options I have here so I'll just click run again and I could see here that I could select other configuration options for example I could increase this to 500 virtual users and to increase the test duration to two minutes and I could make it ramp up in the first two minutes foreign