Regression in r18

[WisdomWolf]

There appears to have been a significant regression in how system and application names are reported to Papertrail in r18. In r17 the system and application name were properly broken out to show the node that the pod was running on and the pod name for proper filtering, but in r18 it now has generic unhelpful designations. This screenshot shows log entries from r18 and then r17 when I reverted my deployment (changing nothing else about the deployment other than the image tag).

[girishranganathan]

@WisdomWolf, thx for pointing out the differences.
I have taken a stab at a fix for this issue. Can you please give this image a try: quay.io/solarwinds/rkubelog:r19-b2

[WisdomWolf]

Looks good, thanks!

[tuttle]

I have the similar problem: This is what I see on Papertrail's Events view:

Using quay.io/solarwinds/rkubelog:r18:

Apr 03 00:29:07 222.111.222.111 logger mysystem-client-k1/application rkubelog/node(macek)/pod(client-k1-application-8qji33ix-75fddb4886-g4pgt) - 10.19.0.1 - - [03/Apr/2021:00:29:07 +0200] "GET /app-liveness HTTP/1.1" 200 2 "-" "kube-probe/1.19"

Using quay.io/solarwinds/rkubelog:r19-b2 and quay.io/solarwinds/rkubelog:r19rc2:

Apr 03 00:44:07 mysystem-client-k1/application rkubelog/node(macek)/pod(client-k1-application-8qji33ix-75fddb4886-g4pgt) 10.19.0.1 - - [03/Apr/2021:00:44:07 +0200] "GET /app-liveness HTTP/1.1" 200 2 "-" "kube-probe/1.19"

IOW, the public IP address of the cluster is gone, also the word logger. Unfortunately, the IP address is useful for me for classification.

[girishranganathan]

@tuttle I will look into this. Thx for bringing it to my attention.

[girishranganathan]

hello @tuttle, I have taken a stab at this issue. can you please try r19rc3 image?

[cfroystad]

@girishranganathan Unfortunately, we're still seeing the issue pointed out by tuttle in 19rc3