-
Task
-
Resolution: Done
-
Major
-
None
-
None
-
5
-
False
-
False
-
NEW
-
NEW
-
-
-
Logging (Core) - Sprint 199, Logging (Core) - Sprint 200, Logging (Core) - Sprint 201, Logging (Core) - Sprint 202, Logging (Core) - Sprint 203, Logging (Core) - Sprint 204
Summary
Patch the fluent collector image with changes to the in_tail and prometheus plugin to expose the metrics needed to understand the volume of logs being collected.
Acceptance Criteria
- Add patch to origin-aggregated-logging/fluentd for the changes needed to modify fluentd
- Create PR in upstream fluentd and prometheus plugin to land changes in the original code base
Notes
- is blocked by
-
LOG-1623 Metric`log_collected_bytes_total` is not exposed
-
- Closed
-
- is documented by
-
-
- Closed
-
- is related to
-
-
- Closed
-
- links to
[LOG-1213] Productize collector metrics changes for fluentd
Hi pmoogi, I find another issue relates to this, PTAL at https://issues.redhat.com/browse/LOG-1647 .
Hey pmoogi , I drafted one TC[1] , could you please help to review it? You can find the username/password from [2]. If you have any comments, you can add here or polarion, if the case looks good, please add `tc-approved` label to this task. Thanks.
[1] https://polarion.engineering.redhat.com/polarion/#/project/OSE/workitem?id=OCP-43177
pmoogi jcantril@redhat.com, I opened a bug here: https://issues.redhat.com/browse/LOG-1623
qitang@redhat.com please create a bug associated with the epic that pmoogi can link her PR to. I want to ensure we have proper tracking
qitang@redhat.com I have verified that it is a fluent.conf problem. I am pushing a new PR to generate new fluent.conf which will include fluentd new plugin loading. Please wait until today to get that merged. 
> for fluentd to publish log_collected_bytes_total metric via new plugin, we need to change the fluent.conf
pmoogi To do this, we must set the managementState to `Unmanaged` or the changes will be overwritten by CLO, I think this action is dangerous for customers.
Besides. I set `managementState: Unmanaged`in cl/instance, then added following lines to fluent.conf and restarted all the fluentd pods, I still couldn't get the metric `log_collected_bytes_total`:
<source> @type collected_tail_monitor </source>
jcantril@redhat.com do we need to change ./pkg/generators/forwarding/fluentd/templates.go to add below source plugin -
@type collected_tail_monitor
<metric>
name log_collected_bytes_total
type counter
desc Total bytes collected from file
<labels>
tag ${tag}
hostname ${hostname}
</labels>
</metric>
for fluentd to publish log_collected_bytes_total metric via new plugin, we need to change the fluent.conf
to add the below lines:
<source>
@type collected_tail_monitor
</source>
I checked the logs in fluentd container and logfilesmetricexporter container, no errors.
workflow didn't correctly change, fixing affected issues