Evaluating Elastic APM for instrumentation of our Java web services. Viewing a transaction sample only displays spans of type DB. What other types should be displayed without manual instrumentation via e.g. Span API? Cache, application, external? Looking at demo application none other than DB span types are displayed.
Future plans for this?
Thanks for evaluating our solution
Please take a look at the supported technologies page to see what's currently supported automatically.
We will love to know what you are using and miss the most, here in the topic or by taking this short survey.
Clearly there are plans for more automatic support, but we will love any input in order to help us prioritise.
That link didn’t answer my question. Given that the Java application is fully supported technology wise; what span types are automatically available? DB, cache, application...?
Hämta Outlook för iOS
Well, as you can see in this page, span types would be either DB or HTTP exits. No cache at the moment. I am not sure what
application type means, but if it is your own business code, then the way to go is using our public APIs to monitor those manually.
I hope this does answer the question, let me know if not.
I've played around with it a lot for a few weeks now.
It basically supports Application, usually of the form Controller#method (servlet, RestTemplate) and JDBC calls.
No auto-instrumentation support for JMS, though our company's team is doing some direct agent modification to propagate trace id across a JMS process by mimicking RestTemplate headers.
Thanks a lot for elaborating.
We will love to see what you did for JMS, that is- if you can/want to share, maybe even contribute to the project . Are you adding the trace context as JMS Message property?
Yes that is right, we used JMS message property to propagate trace and are seeing the linked spans under a single trace. There seems to be an issue with the timing numbers though, which we suspect is probably some sync issues in the time stamps between producer and consumer services. I'll provide more info once we get better results and our plan is to contribute that back if it works well for us
That's great to hear (actually read )!
Please don't hesitate to contact us. We can try and assist as much as possible, even for this timing issue.
Feel free to open an issue in our repo or even fork and pull request if there's specific code to discuss.
Thanks and good luck!
This topic was automatically closed 20 days after the last reply. New replies are no longer allowed.