periodic snapshot of table by default happens every 10 seconds. CPU usage shoots up and also each of these cycle takes over 7 seconds.
On local machines this is lot higher.
2019-04-29 10:44:14.921 [info] <0.771.0>@imem_snap:handle_info:310 [IMEM] {imem_snap,310} !!!!!!! Time for snapshot is : 6.989392
2019-04-29 10:44:31.926 [info] <0.771.0>@imem_snap:handle_info:310 [IMEM] {imem_snap,310} !!!!!!! Time for snapshot is : 7.006396
2019-04-29 10:44:48.962 [info] <0.771.0>@imem_snap:handle_info:310 [IMEM] {imem_snap,310} !!!!!!! Time for snapshot is : 7.021399
2019-04-29 10:45:05.997 [info] <0.771.0>@imem_snap:handle_info:310 [IMEM] {imem_snap,310} !!!!!!! Time for snapshot is : 7.021398
2019-04-29 10:45:23.205 [info] <0.771.0>@imem_snap:handle_info:310 [IMEM] {imem_snap,310} !!!!!!! Time for snapshot is : 7.208436
2019-04-29 10:45:40.287 [info] <0.771.0>@imem_snap:handle_info:310 [IMEM] {imem_snap,310} !!!!!!! Time for snapshot is : 7.068408
periodic snapshot of table by default happens every 10 seconds. CPU usage shoots up and also each of these cycle takes over 7 seconds.
On local machines this is lot higher.