Performance Metrics Calculator

This tool calculates the following metrics

• Number of requests sent
• Fastest Response Time in ms
• Slowest Response Time in ms
• Mean Response Time
• Median Response Time
• The percentage of requests that succeeded
• Any error codes returned that weren't a success
• The size in bytes of the smallest response
• The size in bytes of the largest response

---

Usage

1. Install Go
2. Make sure the GOROOT and GOPATH are set correctly
3. Install the following packages (not sure if Go automatically downloads the libraries)
   • "github.com/galdor/go-cmdline"
   • "github.com/montanaflynn/stats"
   • "github.com/sirupsen/logrus"
4. cd into the repo/src
5. Commands
   • go run . --help
   • go run . --url {url}
   • go run . --url {url} --profile {Number of requests}

---

Working

• Definition of Response Time for this tool: The time taken to establish the connection to the server and receive the all the data sent by the server. Hence, if the data sends a lot of data, the response time will increase.
• If there are no successful responses from the server, the Fastest Response Time, Slowest Response Time, Mean Response Time, Median Response Time, Size of the smallest response, Size of the largest response are for the responses with status codes other than OK (200).
• The tool automatically reads the protocol from the URL and forms a socket connection with ssl.
• Currently supports only http and https.
• The connection request times out after 7s.

---

Screenshots

---

Comparison

Number of Requests	Amazon	Google	Worker
100
1000

• For 100,

  • The worker has the fastest time: It served the client with a response in a whooping 76 ms. Amazon comes in 2nd with 88ms.
  • The worker is ~200ms faster than the slowest response time(from Amazon)
  • The average Response time for the worker is fastest and is ~30 ms faster than Amazon (who comes in 2nd)
  • The median time for worker is 97.5 ms. Which means there were 50 requests faster than 97.5ms. Amazon comes in a close second with a median time of 105ms.
  • Overall the worker serves faster than Amazon and Google. While Amazon performs well, there's a clear difference between Worker and Amazon.
  • Since the response times take into factor the size of the response, a better idea could be given by taking the size into consideration. A unit of byte/ms would give the ideal performance of each server.

• For 1000,

  • Amazon has the fastest time for 1000 requests. While it has the largest byte size array, it takes the least time.
  • Amazon also has the slowest time. Since one of the responses was large, it must have taken more time. Ideally, plotting a graph would give more insight into the performance
  • Worker has the least mean time. Which means on an average the worker performs better.