How to Monitor Web Page Serve Time with the SevOne PLA
There are many interesting use cases for the SevOne Performance Log Appliance (PLA), but here’s one we recently uncovered: monitoring the time it takes to serve up a web page using Apache logs.
As Infrastructure Developer Marius Ducea pointed out in his blog Apache Logs: How Long Does It Take to Serve a Request, you can configure Apache logs to report on the time taken to serve a page. This value can be represented in seconds (%T) or microseconds (%D) in the LogFormat line. See the image below for an example:
Once you’ve ingested this data into the SevOne PLA, you can do some interesting things to monitor the performance of your web applications:
- Automatically baseline the metric so that you may receive alerts any time the performance deviates from expected normal behavior (i.e. it’s taking much longer than typical to serve a page)
- Categorize all results by time frame in order to pinpoint problem servers, such as all servers taking longer than x seconds to serve a page
- Compare metrics from internal and external users accessing the site to determine if one group suffers longer delay
- View which URLs are taking the longest to respond
- Compare serve time metrics to the underlying infrastructure metrics collected by SevOne (e.g. CPU, memory) to better understand event correlation
With this capability in place, you will have a solid metric for monitoring the performance of your internal and customer-facing web application delivery.
For more information on how SevOne is advancing the use of log analytics for performance monitoring, visit our Performance Log Analytics page.