Added support for persistent buffering in receiver/gateway #1342
Conversation
|
Could also add separate settings for logs and metrics, something like this: |
helm-charts/splunk-otel-collector/templates/config/_otel-k8s-cluster-receiver-config.tpl
Outdated
Show resolved
Hide resolved
helm-charts/splunk-otel-collector/templates/config/_otel-k8s-cluster-receiver-config.tpl
Outdated
Show resolved
Hide resolved
| setfacl -n -Rm d:m::rx,m::rx,d:g:{{ $clusterReceiver.securityContext.runAsGroup | default 999 }}:rx,g:{{ $clusterReceiver.securityContext.runAsGroup | default 999 }}:rx {{ .Values.splunkPlatform.sendingQueue.persistentQueue.storagePath }}/receiver; | ||
| {{- end }}'] | ||
| securityContext: | ||
| runAsUser: 0 |
There was a problem hiding this comment.
Nit: We probably need to document this runAsUser 0 and above used permissions in https://github.com/signalfx/splunk-otel-collector-chart/blob/main/docs/advanced-configuration.md#data-persistence as well as customer documentation. This would be done for the agent, cluster receiver, and gateway. A number of customers request information like this.
There was a problem hiding this comment.
There was a problem hiding this comment.
I'd suggest adding documentation primarily in the Advanced Configuration: Data Persistence section. Then, include a brief note with a hyperlink in the Running the Container in Non-Root User Mode section that points to the data persistence section.
| @@ -104,6 +104,31 @@ spec: | |||
| mountPath: /splunk-messages | |||
| - mountPath: /conf | |||
| name: collector-configmap | |||
| - name: patch-log-dirs | |||
There was a problem hiding this comment.
Nit: We might be outgrowing "log" functionality here for patch-log-dirs. If it works for now I'm fine with it. It might be time to refactor the names and images a little to make them more abstract to better address these new use cases.
There was a problem hiding this comment.
Yeah, something like patch-dirs would work for all configurations I guess?
There was a problem hiding this comment.
Yeah I'd go with patch-dirs.
There was a problem hiding this comment.
Does persistent buffering require the .Values.logsCollection.checkpointPath directory for checkpoints? I don't think it does. I recommend removing the code that sets up the logsCollection checkpoint directory for the cluster receiver and gateway to avoid potential race conditions. Since the agent is deployed to all nodes and the cluster receiver and gateway are only deployed to some, they may attempt to set permissions on the same directory simultaneously.
|
Please update the docs at https://github.com/signalfx/splunk-otel-collector-chart/blob/main/docs/advanced-configuration.md#data-persistence for these changes as well. |
…luster-receiver-config.tpl Co-authored-by: jvoravong <47871238+jvoravong@users.noreply.github.com>
…luster-receiver-config.tpl Co-authored-by: jvoravong <47871238+jvoravong@users.noreply.github.com>
POC - Added support for persistent buffering in receiver/gateway