why the Latency is the 50+ us , but not the third span end time 164us,
I'm more concerned about the longest Latency 。
Hi!
I had this problem when server sent custom span-logs to APM server without @timestamp attribute (in this case APM Server insert @timestamp to logs when APM Server already accepted the logs).
But I think that the real reason for this problem is the parent-transaction side has different server time then the child-transaction side
This seems related to asynchronous execution - the MarketData transaction handles a request by creating a MarketEvent and dispatching it to an asynchronous handling by some other thread pool and then returns. We consider the duration of the request handling from the perspective of the caller/client, which is the transaction duration (time to respond to caller).
If this properly reflects your application, then I think you shouldn't mess with it, regardless if these are custom spans, like @Spank321 suggests, or automatically captured ones.
What you can do is create your own "Full Market Event handling duration" dashboard based on the raw data- you can look for all spans with a given transaction.id and consider then greatest timestamp.us + span.duration.us as the end timestamp of event handling.
I hope this helps...
Thank you for advice. i trid to create the "Full Market Event handling duration" dashboard ,but faild.How to visualize timestamp.us +span.duration.us ?
I am new to kibana,can you share me a guide to creat the dashboard to show distribution of latency of "Full Market Event handling duration" ?
Thanks again
This topic was automatically closed 20 days after the last reply. New replies are no longer allowed.