Answer by Andreas G. ·
Here is also a related forum post where i gave a brief explanation on why you may see a high number of skipped events: Many skipped events in dynaTrace 6.0 and SocketTimedOutException for FrontEnd server
Answer by Anton G. ·
The best way is to go to the "Agents Overview" Dashboard and look at the Skipped Events Column (the last one).
I checked Agent Overview - the column "Skipped Events" showing '0'. dynaTrace Server Health dashboard shows lots of "Skipped Event (Server)". How could I get detail of "Skipped Event (Server)"?
For Server Skipped Events it is most likele PurePaths that are exceeding their Size Limit (Default = 100000). So - this are PurePaths where the agents keep sending events because they are still ongoing but the server skipps these events because the server already "terminated/truncated/closed" the PurePath because it ran outside the default node limit. Examplese of these PurePaths are very long running batch jobs.
Please have a look at your PurePath Dashlet and see whether you have PurePAths with a Size of 100000. These are the ones that should also be marked as Incomplete/Truncated (they have a grey status icon). If you have these types of PurePaths you need to figure out whether
a) you really need these transactions to be recorded (e.g: exclude the URLs, exclude your custom entry points) or
b) whether you can use a different entry point in order to split the long running PurePath into smaller ones.
Andi
It's confused. I was told by one of consultant who visited my company recently, he said Skipped Events in the Server Helath dashboard were anything except purepath. Skipped Purepath dashlet is the purepach been skippped. Now you said Skipped Event is the purepath exceeding large size. Then what does "Skipped Purepath" tell me?
There are several reasons why the server may skip events. I just wanted to highlight one reason that I see very often being the root cause when we see a very high number of skipped events on the server. And this scenario is caused by PurePaths that run into either a timeout or are truncated because they exceed the purepath node limit. in this case the agents still send their events to the server because they dont know that the PurePath has already been truncated. But as the server doesnt handle these events anymore it simply skips them.
As I said - there are also other cases where you may see skipped events but this is the case I see most often. Best way to verify this is by checking if you have PurePaths that time out or that are truncated
New situation, got a support request in...
WE need to set the Size above 10,000 due to long-running batch processes, and you didn't tell us how.
We have known data sizes of 30,000 and 100,000 records, which will definitely run for long duration. Further, app architecture appears to start processes, then leave them running (like Daemons).
Suggestions? If I get anything from the support ticket, I'll place the follow-up here.
Hi
it would help to see a PurePath and what you mean with "30000" records. If records means items in the Database or SQL Calls to a database then this is not a problem as dynatrace is aggregating DB Calls - unless these are 30k different SQL statements.
In general we advice that long running processes (daemons) should not start a PurePath when they start their main routine. You should find a better entry point such as a "processItem" or "processMessage". In that case you will end up having several PurePaths that indvidually finish in a reasonable amount of time and will not exceed the limits. If you are able to capture some unique JOB Id or whatever it is that uniquely identifies that Batch Job Run you can create a Business Transaction so that you can easily look at all PurePaths that were captured PER JobID
Andi
JANUARY 15, 3:00 PM GMT / 10:00 AM ET