The Log Monitoring -> Events page at Dynatrace Managed shows this notification:
"Log events will be deprecated and replaced by custom metrics. For now you can still use events."
On the other hand, the documentation for custom metrics for log monitoring states "A custom metric can be created only for monitored logs (log files that are stored on the Dynatrace server)":
So essentially does this mean that if you're not using the centralized storage for log analytics, you will not be able to create alerts based on logs anymore? If yes, when will this change take place, i.e. the whole Log Events section will not be available at Dynatrace Managed anymore?
Solved! Go to Solution.
Hey Kalle! Log Monitoring Product Manager here. I fully understand your concerns regarding this change however this decision was made solely based on customer feedback. We want to generate events base on streamed log data (aka centrally stored) because many customers want to benefit from this feature for different log sources (logs from cloud services, fluentd, syslog etc.). Right now we are working on functionality that will allow you to parse log message and create metric based on parsed content.
From product and licensing perspective we will be also evaluating possibility to drop log records from log stream after metric/event creation - this should address setup and cost concerns. Happy to discuss this offline.
Hi Michal, thanks for the heads up. I can of course appreciate that for bigger log analytics needs, the new features are very much needed. I suppose the requirements in my customer cases so far have been pretty limited; just some individual log alerts where the log events were actually a quite nice fit. For data visualization, the metrics generation is of course vital to have.
I have basically exactly the same problem. Log storage is not that valuable to us (due to absence of containers) but log events are very useful whereas I'm not sure the client is going to be willing to pay for full blown log monitoring. The free tier is super small and even just monitoring three process logs, one had some issues and used up the quota all of a sudden and then seems to make log monitoring and/or metrics unusable from that point on.
To me it is still unclear why removal of one is needed for the other to be implemented. I can understand the argumentation for adding the new feature, but I don`t understand the argumentation for removing "detection rule for custom log events". That functionality was critical for customers, so the argumentation for removing it is hard to understand.
Please elaborate on why Dynatrace is removing ease-of-use functionality.