Unknown content monitoring

Realize the simplest host survival monitoring.
Each node starts a monitor service, and gives him a list of monitor services of other hosts (IP + port number).
After the monitor is started, every timeVisit the monitor service of all other nodes once, and the monitor service of each node is the same.
Next, the monitor sorts the corresponding time of each node, and checks whether there is no corresponding one, or the corresponding one is too slow.
ThenNext, each node synchronizes information with each other, and checks whether some nodes do not correspond to others. If most nodes find that a node cannot respond, then the node is considered abnormal. Send an alarm together.

In this monitor cluster, each machine deploys a service, does not use the master node and does not consider the problem of master selection. As long as the number of services specified in the cluster accesses a node and returns an exception, then it is considered that the other party has a problem

p>