Please see http://www.zabbix.com/ for detailed information about Zabbix.
On-line Zabbix documentation is available at http://www.zabbix.com/documentation/2.0/manual/installation/install .

Features added:
- push metrics to zeromq queue (queue is expected to bind PULL socket - pipeline pattern)
- time fix switch
- housekeeper optimization
- store history in zabbix sql only for items with triggers (used when data is sent to queue)
- disable storing trends in zabbix sql (used when data is sent to queue)
- added availability computation
- performance optimizations

Config options added:
ZmqQueueAddress - zeromq pull socket address for metrics (possible to specify a list to distribute
    metrics to multiple queues) eg: tcp://localhost:9901,tcp://localhost:5501
ZmqErrQueueAddress - zeromq pull socket address for errors related to pushing to metrics queue
    eg: tcp://127.0.0.1:8801
ZmqQueueRecoveryDir - dir where zabbix can store recovery info about metrics pushed to queue,
    useful when queue was down and metrics must be resend
ZmqQueueDaoc - use sock opt ZMQ_DELAY_ATTACH_ON_CONNECT
TimeFix - whether zabbix timefix is enabled or not
HousekeeperSingleQuery - should housekeepr use single query
HousekeeperSleep - how long house keeper should sleep between batches it processes
TriggersHisotry - store history only for items used by triggers
TrendsSqlWrite - whether to write trends to zabbix sql or not
AutocreateLimit - limit for simultaneous autocreations of metrics
TrapperProcessingLimit - limit of how many items trapper can process in single message
AnalyzerAvailDir - directory where analyzer stores periodic availability snapshots
AnalyzerAvailInterval1 - length in seconds of first interval computed by analyzer
AnalyzerAvailInterval2 - length in seconds of second interval computed by analyzer
AnalyzerAvailStoreInterval - interval for storing periodic snapshots of availability (seconds)
AnalyzerAvailPingFreq - lenght in seconds of maximum gap between two datapoints, if gap is longer
    analyzer starts considering particular metric as showing unavailability
TargetTrends - tag marking trends sent to queue
TargetAvail - tag marking availability sent to queue

Docker:
docker build -t zabbix .

docker run --name mysql0 -v /opt/docker/mysql/data:/var/lib/mysql -e MYSQL_ROOT_PASSWORD=docker123 -p 127.0.0.1:3306:3306 -d mysql:5.6

docker run --name zabbix0 --link mysql0:mysql -p 8443:8443 -e ROUTER_QUEUE_ADDR=tcp://127.0.0.1:9901,tcp://127.0.0.1:9903 -e ERROR_QUEUE_ADDR=tcp://127.0.0.1:8801 -d zabbix:latest

docker run --link mysql0:mysql -e QUEUE_ADDR=tcp://127.0.0.1:9901,tcp://127.0.0.1:9903 -e ERROR_QUEUE_ADDR=tcp://127.0.0.1:8801 -it --rm zabbix:latest /bin/bash