-
Bug
-
Resolution: Unresolved
-
Minor
-
None
-
rhos-18.0.z
-
False
-
-
False
-
?
-
rhos-ops-platform-services-pidone
-
None
-
-
-
-
Sprint 2, Sprint 4
-
2
-
Important
To Reproduce Steps to reproduce the behavior:
Default Memory high watermark setting for RabbitMQ is 0.4 of available memory. This doesn't make much sense for RHOSO scenario where RabbitMQ seem to be using isolated RAM and there is no point in such aggressive limitation for it. Please consider changing this limit to higher default value.
Expected behavior
RabbitMQ operator will enforce reasonable default watermark settings that will work as expected in RHOSO scenario where dedicated RAM is allocated to RabbitMQ pods.
Bug impact
This situation is a good candidate for root cause behind sensitive situation in customer's environment where RabbitMQ started blocking connections while load was relatively low.
Known workaround
https://issues.redhat.com/browse/OSPRH-16586 blocking straight workaround, it is possible to bump pod's RAM, or to use rabbitmqctl to set higher watermark.
Additional context
must-gather and rabbitmqctl outputs are attached to support case
- relates to
-
OSPRH-16586 additionalConfig blocks rabbitmq pod start if defined
-
- Closed
-