- ship receive timestamp with internal cns messages and option
for cns worker to discard old messages
- bug fix for memory leak in jmx monitoring bean for cns worker
- accept access key in http header (aws4 style)
- support for WaitTime parameter in cqs ReceiveMessage() ("long
polling") - implementation based on servlet api 3.0 async request
handlers
- added data center flag to heart beat tables and display in
dashboard screens - used by long polling feature to ensure data
center locality
- added redis clear cache api and added more monitoring to jmx and
dashboard screens
- deepened health check api to check redis cache, cassandra ring
and cns worker state
- bug fix for jetty http header parsing
- bug fix improving performance by removing legacy code to auto-migrate
old-style cqs queues
- maintain and display basic api call stats in dashboard screens
- allow multiple api endpoints on same host
- added pagination etc. to navigate large number of queues, topics,
subscriptions in web ui
- no need to switch hosts any more with absolute urls when switching
between cns and cqs in web ui
- cleaned up and improved performance of xml responses for all
api calls
- performance tweaks for fast queue and topic creation specifically
for large numbers of queues and topics
- increased default in-memory cache sizes to 1 mio
- harmonized and cleaned up log4j logging
- cns workers will bootstrap and continue to function even if cqs is
(temporarily) unavailable
- bug fix for listing queues by prefix api
- support for DelaySeconds parameter in cqs SendMessage()
- ability to trace messages through logs by receipt handle