| 27 Jun 2023 |
|  @martin.mueller:dataport.modular.im changed their display name from Martin Müller (Phoenix AMO - int.) to 🐦Martin 🔥Müller (Phoenix AMO). | 12:57:59 |
 pepov | No because namespace labels are not attached to the log as metadata, you can only work with pod labels. | 17:18:16 |
| pepov | it uses the same securitycontext as the fluentd pods (under the logging resource: spec.fluentd.security.securityContext) | 17:20:28 |
| pepov | if you have a policy engine you could try to add namespace labels to pods dynamically as they get created (for your previous question) | 17:21:15 |
| pepov | which output are you using exactly? | 17:21:50 |
| pepov | Hey folks Just recently started looking | 17:28:08 |
| 28 Jun 2023 |
|  alltilla1337 changed their display name from alltilla to alltilla1337#0. | 07:42:00 |
| alltilla1337 changed their display name from alltilla1337#0 to alltilla1337. | 07:42:03 |
| 3 Jul 2023 |
|  johanot#6347 joined the room. | 08:43:25 |
| johanot#6347 | the log forwarder (fluentd in my case) is a single-replica statefulset. Why? I know I would (maybe) risk duplicate delivery entries if I let it surge. | 08:43:26 |
| johanot#6347 | asking because: just had a hard node failure overnight and fluentd got stuck in Terminating (apiserver didn't know the status of it, understandably). If fluentd had had surge allowed, a new fluentd could have been spawned on another node while I was sleeping. AFAIC that wouldn't have been dangerous. | 08:48:07 |
| pepov | the log forwarder fluentd in my case is | 10:54:11 |
| 4 Jul 2023 |
|  lazarimre78 joined the room. | 07:17:05 |
|  YourMomIsNotMale joined the room. | 11:21:01 |
|  therealrandomstring joined the room. | 11:27:06 |
|  _ollster changed their display name from MaxR#2646 to _ollster#0. | 13:15:54 |
| _ollster changed their display name from _ollster#0 to _ollster. | 13:15:55 |
| _ollster | hey folks, im currently struggeling a bit to configure kafka outputs
looking at the CRDs im currently not able to directly set a topic or a key for the Certificate im using for auth, correct? (https://doc.crds.dev/github.com/kube-logging/logging-operator/logging.banzaicloud.io/ClusterOutput/v1beta1@4.2.0#spec-kafka ) | 13:15:55 |
| 5 Jul 2023 |
|  alexb271 joined the room. | 15:51:33 |
| 6 Jul 2023 |
|  bazsi77 changed their display name from bazsi77 to bazsi77#0. | 12:39:55 |
| bazsi77 changed their display name from bazsi77#0 to bazsi77. | 12:39:56 |
| 7 Jul 2023 |
|  hofione changed their display name from HoFi to hofione#0. | 14:48:25 |
| hofione changed their display name from hofione#0 to hofione. | 14:48:26 |
| 8 Jul 2023 |
|  cpuram1 joined the room. | 20:20:08 |
| 10 Jul 2023 |
|  bruno.bressi changed their display name from bruno.bressi to bruno.bressi#0. | 15:01:51 |
| bruno.bressi changed their display name from bruno.bressi#0 to bruno.bressi. | 15:02:20 |
| bruno.bressi | Hello there 👋 We're currently trying to figure out a multi-tenancy solution with the logging operator. We aim to make the fluentbit and fluentd more robust, when a user provides an invalid config (like missing secrets / connection failures to outputs). For example, when a user has a faulty configuration, the logs are "stuck" in fluentd and won't reach their destinations or get stuck in the fluentd buffers. Any suggestions or general guidance? We're running multiple multi-tenant clusters, in which users may only use Flows and Outputs and the logging stack (fluentbit, fluentd) is completely transparent to them. | 15:02:21 |
| 11 Jul 2023 |
| pepov | Hello there 👋 We re currently trying to | 08:34:08 |
|  josephsaber joined the room. | 08:50:11 |
|  mranno changed their display name from Anno to mranno#0. | 11:57:42 |
