Performance: Interpreting Trends

Websites exhibit trends that provide performance monitoring gold.  Right now, the response time of a website is different than it was 5 minutes ago, 5 days ago, and even 5 years ago.  Because of these differences - no matter how small – a trend will emerge to offer some interesting results.

There are four important trends to look for when analyzing response time performance data:

1.) Hour of the day, showing the slowest performing and fastest performing hour(s) of the day.

2.) Day of the week, showing the slowest performing and fastest performing day(s) of the week.

3.) Day of the month, showing the slowest performing and fastest performing day(s) of the month.

4.) Day of the year, showing the slowest performing and fastest performing day(s) of the year. 

(A fifth trend, month of the year, can be important to businesses that operate on a seasonal basis.)

For an online retailer, the fastest performing response times could be during nighttime hours and the slowest performing response times could be during daylight hours.  Response times with expected fluctuations may indicate that the web server is being taxed during busy periods in the day and may suggest that load balancing could alleviate the burden.  A flat line trend with little variation but a high response time may indicate that the server is sized too small, needs optimization, or needs help with load balancing.