26 Mar 2023 12:03 PM
The other day, I got to see an interesting message from the Linux kernel. Indeed, it's last words:
In this particular configuration, I can't have Log ingestion, because it's way too much data. Any idea, on alternative ways to get Linux kernel messages, without log ingestion? I'm also thinking of an OneAgent extension, but are there other ways?
26 Mar 2023 12:24 PM - edited 26 Mar 2023 12:25 PM
Hi,
Maybe Metric scripting integration but I am not sure if this is best option for your scenario.
Best regards
26 Mar 2023 03:21 PM
@AntonPineiro, to do that, I would need another program running to detect & ingest. OneAgent extension might be easier to deploy & manage.
27 Mar 2023 08:32 AM
Hi,
Yes, no more ideas come to my mind, a script running locally or OneAgent extension.
Best regards
27 Mar 2023 04:52 PM - edited 27 Mar 2023 04:54 PM
Look Antonio,
probably you are already aware of it but I really like this feature of Oneagent (the scripting integration):
https://www.dynatrace.com/support/help/shortlink/pipe-metric-ingestion#piping-process-output
you pipe the stuff that you like from unix command line to the dynatrace_ingest tool. It is sent on localhost:14499, catched by Oneagent and sent it back to your cluster.
In theory it will scale your DDU as a custom metric (but I need to better verify that) and if this is the case you could spend less than with log monitoring since some DDU are included when you have already HostUnit bought and Oneagent installed.
Downside is of course that you need to sweat more for a solution and there will be a custom piece of something inside your machine.
Cheers
28 Mar 2023 12:59 PM
Thanks for your reply. I could even try to ingest with logs, but the main issue here is it the Kernel... Running another process might not be desirable, might try the OneAgent extension. Thanks!