The new feature "Configurable queue threshold alerts" requires documentation in the user guide as to how to configure the thresholds and how to receive and interpret the resulting events.
Alison, another one for you to qa_ack
The QA review: - There is need to highlight that qpid.alert_count and x-qpid-maximum-message-count are equivalent for setting message count alert threshold - There is need to highlight that qpid.alert_size and x-qpid-maximum-message-size are equivalent for setting byte count alert threshold - Similarly for qpid.alert_repeat_gap and x-qpid-minimum-alert-repeat-gap pair - There is need to highlight that if both message count and byte count alerts are configured then single alert will be triggered for condition which occurs first - reference to self.do_threshold_test() is useless for customers. I would rather refer to whole unit test, for instance: Example of queue alerts usage can be found in: # rpm -ql qpid-tests | grep threshold.py$ /usr/lib/python2.4/site-packages/qpid_tests/broker_0_10/threshold.py The other way which might be more useful for customers can be Gordon's example mentioned in bug 660289, comment 5 pont 2] - this uses just well known tools qpid-config & spout and drain. -> ASSIGNED
I recommend to better specify the qpid.alert_repeat_gap / x-qpid-minimum-alert-repeat-gap functionality according bug 660289, comment 12. Especially the fact that there is no mechanism for delaying the queue alert somewhere and retriggering automatically after qpid.alert_repeat_gap / x-qpid-minimum-alert-repeat-gap period. i.e. having queue in state which caused first initial alert is not enough for seeing the next alert (after lert_repeat_gap period) automatically - there has to be queue state change...
I'm fine with the feature description. -> VERIFIED