The performance test metrics are plotted in one-minute increments. All of the load generating servers feed data back to the LoadStorm reporting engine. 
Click on image to see full size.
Volume graphs show the traffic measurements between LoadStorm and the target web application. Basically, how strong is the storm?
The Summary graph shows high-level metrics of how the target web application performed. It includes a combination of performance metrics.
Summary metrics:
- Requests per Second
- Concurrent Users
- Peak Response Time – Time to get slowest resource
- Throughput – Kilobytes per Second
- Errors
- Average Response Time – Average time to get site’s resources
Click on image to see full size.
Response Time Distribution graphs show the spread of response time data in a graph. This is the range of wait times that your end user could be experiencing. Use this graph to sort out your outliers from your true test data.
Response metrics:
- Min/Max – Highest and lowest response times at that point of the test
- 90% (1.28 σ) – Response times within the 90th percentile (1.28 standard deviations)
- 68% (1 σ) – Response times within the 68th percentile (1 standard deviation)
- Average (μ)
Click on image to see full size.
Finally, Error graphs show the percentage of errors reported during a test. These are the types of problems the VUsers are experiencing. Find out when certain error types become a problem as your site scales. Below are just a handful of the errors we catch.
Error Types
- 401: Unauthorized – A VUser failed authentication to access the target server.
- 403: Forbidden – A VUser couldn’t get to a resource.
- 404: Not Found – Usually the link is invalid due to a typo, permissions, or the file does not exist.
- 500: Internal Server Error – Likely an issue with your server. Does it look like it’s about to crash?
- 502: Bad Gateway – Issue with target servers.
- 503: Service Unavailable – Your server is barely functioning, and can only respond with this status code.
- Request Read Timeout – Resources are taking too long to load, but timeouts are customizable.
- Request Connection Timeout – Our engines can’t establish a connection to the target server.
- Socket Connection Timeout – We barely couldn’t establish a connection to the target server.