How to run a full webpages monitoring mechanism in Oktawave Watch?
Posted by on 12.07.2013 11:49
By analyzing HAR we know exactly which elements contributed to the opening time of the test website and which download phases lasted the longest (divided by querying DNS, establishing a TCP/IP connection, HTTP request sent, awaiting response from the server, receive data). The key is that the analytical data is stored for each of the checks, so we can find out what caused (for example in the previous day) delay in the loading of webpage.
To use this feature, you must log on to the Oktawave panel and from the Services tab go to Oktawave Watch.
Then, in the Services tab, click Add new service. Below is a list of monitored services, if you have previously turned on the monitor for any OCI instance or application.
Select the type of service: FullPage or FullPageHttps, depending on whether you want to check the site with SSL (https).
Then, in the configuration of monitoring at least fill the field for the URL. You can also set many configuration options, from frequency, to match text that should or should not be on the site.
When you set all options, you can save the configuration - then monitoring starts.
On the site with monitored services of watch.oktawave panel newly added service will appear (name is the URL that was entered in the configuration). Monitoring service will take place 24 hours a day at intervals specified in the frequency parameter.
The first results should be within a few minutes of saving the settings. Below we present data collected from a long period of monitoring the service, just to show how to use the analysis of HAR.
From the list of monitored services select a sample service by clicking its name.
This will open a page with a summary of the monitoring and speed charts and also the availability of the selected service. On the following screen, we see (marked on the chart) area that we want to study, because of the increased loading time of the page. Make a selection by clicking on the graph and dragging the selection with the left mouse button.
When you release the mouse button, chart automatically enlarges the selected area. If the approximation is sufficiently large, we will see each test divided into locations where they were executed. Choose the check moment we are interested in by clicking on the corresponding point of the graph (1).
All the checks from this point on will be selected, below the graph name of the location from which the test was performed (2) will appear. Clicking given name, will see a new window with a waterfall chart showing analysis of HAR.
The diagram shows the various elements of the page, HTTP status, the size of the item and the time to download it, all broken down into phases. Hovering the mouse over the time bar displays additional information about the duration of all phases of content download.
It is also possible to view the server response headers for each element by expanding tree and also it is possible to view overall statistics on types of active elements by clicking the Show/Hide statistics button.