We are trying to make application detection work manually. We are confused on how to map entry points/web requests to the application buckets.
Specifically what do we do if multiple web requests from multiple tiers have the same entry point uri (eg. '/')?
Can we configure the application based on pattern of agent names, host names or other fields from entry points/web request dashlet?
For instance, in the attachment, we would like to group the highlighted applications into one.
Have you seen the following doc page that explains how Application detection works? Conceptual Overview
Right now a User Action or a PurePath (if you dont have UEM) will be assinged to a single application based on the pattern matching of the Root URL of this User Action/PurePath. Part of that full URL is the HOST name, e.g: www.mycompany.com. So - you can create apps based on that host name but not a hostname of your internal servers. Not by agent name either or any custom entry points.
There have been requests on extending the Application Detection capability to e.g: have application names assigned to custom entry points, ... - its the first time I hear it that you want to base it on Agent Name or physical host name.
Looking at your screenshot I see you have auto-detection turned on. Have you ever thought of definiging your own custom applications such as "Contains 'subscriber'". This would put all your highlighted Domain names into a single logical application
Thanks for the reply! We did try using a match like "subscriber" but we were getting odd data output. only one of our URI's actually managed to work (one for "reachclient"). We did not do it how the documentation has it in the screenshot. We'll give that a try.
One question: In some cases, we have multiple applications in which the purepath "name" is just a forward slash. These purepaths belong to a number of different apps but are all called the same thing. How would we group those together? Should we just be able to find everything based on the web requests dashlet or based off of application auto detection?
Thanks Andi for the detail information!
However, as Shrimant mentioned, we've tried custom application using "Contains 'subscriber'". Unfortunately it does not capture any of the highlighted application,
The documentation was very helpful - the trick that made things work for us is to add "://" as part of the contains statement to tell we are looking at the root url (...I suppose).
It partially solved our problem when the application that we need to group by is the first part of the domain name (e.g. http://subscriber.xxx.com, http://subscriberxxx.xxx.com can be properly grouped), while for cases like http://xxxsubscriber.com, http://yyysubscriber.com this would not work because simply using "Contains 'subscriber'" does not seem to check the root url.
Do you have any suggestion/comments on what would be a best practice on configuring applications? What was the reason that "Contains 'subscriber'".did not work?
Thanks a lot,
Thanks for the clarification. I wasn't aware that :// needs to be part of it. This doesnt make sense to me. I will forward this to the engineering team and see what they have to say. For me it should work by just specifying ANY part of the full URL. Let me ping the team and I get back to you
I just checked with engineering and also tested this on my local instance. The contains should work as expected. There is no need to have the :// in the pattern. Can you double check on whether you may have an accidential typo?
we tried using part of the URL without the :// for a number of our applications and not a single one was detected correctly. in the page you linked, they started it from :// as well so we tried that and it worked. A typo would be possible, but not 15 times in a row (one for each app)
Thanks for the session. If you have the chance - can you try the following
#1: disable auto-detection of apps
#2: remove the :// in your custom app definitoins
#3: if the problem still happens send over these PurePaths that should then end up in the Default Application
I had the same issue and this is how i managed to resolve it. If you have a web server fronting your applications its easy although you will have to add host capture request attribute to your web server and after that when you create your ppath-bt you will need to ensure that you change your web request filter to capture from entry-point.
make sure that your splittings are correct, in my case i managed to split using the first 3 parts of the URI using this regex using the Web Request - URI Pattern Values
Make sure you also change the Occ on Purepath : entry point.
When you get to your application specify the address then the URI pattern value in this case it would be something like this if.
Once your BT is associated you should have something like this:
Lastly on your web server agent group ensure that you have the web server sensor and make change it to "active and start new purepath" on your application server ensure that you have your servlet sensor and make sure it is configured to only "active"
I have tried manually defining the application by taking entry points from entry point dashlet , used class /method as entry point. Application is working fine and i am getting full details in web ui . Challenge i am facing is we are getting all auto detected apps also in listed in applications by host id / web application id and it is mesing up whole application page . If I turn off auto detect transaction flow is not coming up in web ui. Is there a way we can just have custom applications only listed on application page in web ui ?
In the Web UI there is no way to hide/remove the auto-detected applications.
Just a quick question that if you already created a user defined application then why do you need automatically detected applications identified by Web application Id or Host. I guess you can un-check that option in the applications tab.
Hi @Andreas G., we need to configure Dynatrace Agent to our java batch jobs but we are unable to do. We have tried different entry points, but dynatrace is not at all identifying transactions running for particular batch job.
Alternatively as of now we are capturing batch job stats like method hotspot and execution of all queries with the help of CPU sampling . Is there any guide available with which we can configure dynatrace correctly ?
FOR UI/Portals Dynatrace is working perfectly fine without any issues and but not working for batch jobs