We are working on a product which has several AWS nodes which we need to monitor. Some of these nodes have an Application Server associated with it. But the others are just standalone JVMs. We successfully configured Dynatrace agents on all the required nodes and instrumented them using the Dynatrace Client. But when we view the overall diagram of connected agents (transaction flow), we are unable to see the standalone Java machines. They are visible in the host list. Any clues why these standalone JVMs are not seen in the transaction flows? Due to this we are not able to view any associated PurePaths/method hotspots etc for these machines.
Is there any message that instrumentation is disabled for those agents in the agents overview dashboard?
How the transaction invoke a method from standalone JVMs?
Is servlets sensor is active and start PurePaths?
Auto transactions/PurePaths starts only for the Web Server, Servlets and ASP.NET so might be we will have to instrument the custom methods or extend the properties as receive entry point sensor.
Hi Babar, no such msg in the Agents Overview dash. Could you please let me know where I could check the Servlet Sensor? I am a Dynatrace newbie, would be great if you give me some pointers on custom instrumentation. Appreciate!
Just double click or right click and the system profile and select the "Edit System Profile" then select the "Agent Group" and click on the "Sensor Configuration" so on the right side you will see all the placed sensors.
Among all other you will find a "Servlets" Web Requests (Servlets): Collects the response time and context information of Servlets and Java Server Pages (JSP)." and in the capture column you can see the status of whether it is only active or active and start PurePaths.
Since it is looking at actual traffic the transaction flow will only show the agents/tiers that have been involved in transactions/PurePaths during the timeframe you are viewing. It certainly is possible that it is just that these agents have not been involved in any transactions yet or in the window you're viewing.
I would also check the "Capture" column in the agents overview to ensure that it set to true and take a look at the "Captured PurePaths" column. This will tell you whether DT has created any PurePaths for those agents. If the count is zero then it might need custom instrumentation like Babar mentioned.
The 'Capture' column for the agent shows true, 'Total PurePath Count' is zero. I didn't see the 'Captured PurePaths' column, but there was a 'Completed PurePaths' column which showed zero again. I am a newbie at Dynatrace, will look up the custom instrumentation that you mention.
Is it possible you're missing any tiers in front of those JVMs ? Also you might want to double check your OOB sensors starting transactions (i.e. Servlet sensor set as Active and start Purepath, ...). You might also need to add custom sensor to start purepath, you'll need somebody that knows the code for that.
Thanks a lot Babar, David.
I created custom sensors for the profile and placed them for the required agents. I am now able to see the agents for the standalone JVM nodes in the transaction flow and PurePaths are getting generated for them. The video was a great help.
But after adding these sensors, I see that each time I trigger a load , a new agent (with suffix ) automatically gets created. When the load test is stopped, the new agent goes away.
Anything I am missing here?
Appreciate the help.
The way it works is if 2 agents have same name, Dyanatace will add a suffix. It seems to me your agent is getting disconnected and when it connects back DT sees it as a new one. It's happening to me too.
You don't need to worry about this because this is the default setting by DynaTrace which is configurable in the collector's config file.
The second agent will be appeared with the suffix if we restart the agent and within next 72 hours it will be automatically removed if we do not restart the agent again or it does not restart unexpectedly.
Hi Babar, Stephanie,
This second agent is getting created only for the agent which I configured the custom sensors for. The other agents related to standard App server nodes do not display this behaviour. Any reason that this additional agent shows up especially for the agent with custom sensors? Also, should we be monitoring the original agent along with the  agent, since both show up in the transaction flow? Apologies if the questions are trivial.
Did you restart this agent in near past? Might be you restarted due to unavailability of host sensor. If Yes !
I would say to wait for 72 hours which I already said in my last comments.