Red Hat Bugzilla – Bug 1466494
[dedicated][RFE] Aggregate router logs
Last modified: 2017-09-25 10:16:00 EDT
1. Proposed title of this feature request Aggregate router logs 3. What is the nature and description of the request? We need to aggregate router pod logs. When an outage occurs in a customer cluster, we need this information to solve issues, or at least provide RCA to prevent issues in the future (rather than just waiting for them to occur again). 4. Why does the customer need this? (List the business requirements here) Stability of cluster 5. How would the customer like to achieve this? (List the functional requirements here) operations logging cluster
Good morning! This RFE has been opened for over 4 weeks and has not yet been triaged. (I.e. the pm_ack, devel_ack and qa_ack flags get set.) Is that something that should have happened, by now, for this ticket? Thanks.
Primary cause is being working (alerting ops when kibana web console isn't responding whether it is router issue or other), so this will minimize the need for this feature. Feel free to close if no longer needed. I will check into creating a card on the Trello board to track this but see it as pretty low level priority as ops will be monitoring this themselves. If we find other use cases or customers who want this, it would bump it up a bit. I need someone from ops to highlight how router logs are captured today. Adding bign for that.
+ Wesley to provide the relevant information
As for today they are captured just like every other pod log(through fluentd). They are stored so only cluster admins can see them, which means we have to manually use curl to access the logs, not something OPs has documented. Though even with that I don't think router logs are overly helpful as it is mostly just logging when the router refreshes its cache of routes.
I believe this can be accomplished by getting access view the tier access for SREs and CEE folks. The will give them cluster reader rights to they have access to logs for all pods. See https://mojo.redhat.com/docs/DOC-1144200 I believe tiered access satisfies this RFE and this can be closed.
That would work, agreed, I think.