Nrpe socket timeout


I have 4 linux servers that are monitored by nagios. Sometimes i get message on my contact e-mail:
CHECK_NRPE: Socket timeout after 10 seconds.
Notifications stating that nrpe timed out come for ALL services and for ALL hosts.
Then, i got another message that eveything is fine. This flapping events take place every 1-2-3-4-5 hours randomly.
Nagios and monitoring servers are situated in the same network, thus I haveno intermediary between monitoring servers and nagios.
Can you help me to diagnose what’s wrong?
sry for my english