Business Transactions and Creating Incidents
I posted the below question in the Business Transaction Management Forum but did not receive any feedback, was hoping i might get some input here
I'm just starting to work with setting up business transactions and i wanted to get some feed back on the best practices for setting up a few business transactions. We are transitioning over to dynatrace and we utilizing another APM tool right now to monitor performance. I'm trying to figure out how to create the same type of monitors in DynaTrace. For example we have a .NET monitor that looks at various IIS Application Paths, for example ARG/ARGApi/v3, the current monitor will trigger an alert if the average request time exceeds 1ms, if the % of performance events/ sec and % of exception events/sec exceed 45%.
What i'm wondering is what is the best way to monitor this IIS Application path via DynaTrace, is using a business transaction the best apprach or is there a better way to handle these type of performance metrics?
Any help would be greatly appreciated.
Thanks
Tim
Hi Tim,
Sorry for the delayed response.
The purpose of Business Transactions is to pull data from transactions (PurePaths, User Actions, or Visits). Since what you're looking for is more of a statistic, BTs will not be helpful to you. You have two courses of action, depending on whether you have .NET agents on your IIS server.
- If you do have a .NET agent installed where you want to monitor the performance counter, please edit your System Profile, go to the Measures dialog, and click "Create Measure." In the resulting window you can press Ctrl+F or simply start typing to search, so search for "windows" and you will see two measures. Choose "Custom Windows Performance Counter" as your metric, giving it a meaningful name for the specific metric you want to pull, and adding the Object, Counter, and Instance (where applicable) info in the configuration.
- If you do not have a .NET agent installed on the box you want to monitor, you can use a dynaTrace Monitor, which uses perfmon from a Windows-based collector instance to remotely grab the performance metrics. Learn more here: Windows Performance Monitor
Hope that helps,
Rick B
Thanks for the response Richard, I took a look at your attached hyper link and these appear to be geared more towards Windows Performance, CPU, Memory, etc. What i'm attempting to do is to monitor the average responce time of a particular URI. So for example i have a landing page of /TB/Sale/Paymentlandingpage/, what i want to do is have an incident generated if the average responce time of that web request is over 900ms. Sorry if my first example was not detailed enought, but is this something that can be done?
Sorry for misunderstanding. This is actually even easier, and still won't require a BT.
- Create a measure on the Web Request Time for that URI. In the Web Requests dashlet, you can right-click your desired URI and click "Create Measures", and Time will be selected by default.
- Add an "upper severe" threshold of 900 (ms) to the measure. In the lower portion of the measure configuration dialog you will be able to define thresholds for your new measure
- Create an incident using this measure as a condition. Edit your System Profile, go to the Incidents dialog, and click "Create Incident Rule..." Add your new measure as a condition here. For more information about configuration of the aggregation and Evaluation Timeframe which is appropriate for you, check out the detail in Incidents and Alerting
Rick B
Sure thing. If you want something a little more advanced (but still relatively easy to configure), but smarter, you can also have a look at Baseline and Smart Alerting Explained. This feature allows dynaTrace to learn your transactions and alert you on deviations in median and 90th percentile response times, as well as failure rate.
Rick B
