I have Zabbix 2.0.3 setup with 1 master and 7 child nodes. All servers are running on CentOs with 1G RAM for childs and 4G for master. When I start DM for 4/5 nodes out of 7, node watcher process is under control (utilization below 50%). But when I connect more than 5 child nodes, all on a sudden it jumps to 100% utilization all the time at all child nodes (note: master node "node watcher" process is well below any kind of utilization). As a result, I am facing synchronization problems. I have tried mysql database tuning. But, as I don't know much about tuning its not helping me out. If you could suggest me some possible get away's from this, please do. Is there anyway to increase the number of node_watcher instances like there are for pollers/trappers? My master .cnf and child .cnf files are given below:

Master my.cnf:

innodb_buffer_pool_size=1500M
key_buffer_size=64M
innodb_file_per_table
thread_cache_size=128
query_cache_limit=256M
query_cache_size=256M
tmp_table_size=256M
max_heap_table_size=256M
table_cache=1024
max_connections=100
innodb_flush_log_at_trx_commit=1
join_buffer_size=2M
read_buffer_size=512k
max_allowed_packet = 4M

Child node my.cnf:

innodb_buffer_pool_size=128M
key_buffer_size=4M
thread_stack=32K
innodb_log_file_size=16M
innodb_file_per_table
thread_cache_size=8
query_cache_limit=32M
query_cache_size=32M
tmp_table_size=32M
max_heap_table_size=32M
table_cache=128
max_connections = 75
join_buffer_size=2M
read_buffer_size=128k
max_allowed_packet=4M

If you could please help, that would be great.