When a service or host changes their state frequently, then it is called flapping that may cause lots of problems and generate too many recovery notifications. Flapping is detected in the following manner:
- Store the results of last 21 checks and then analyze this historical check result to know the number of transitions that are being taken place by the host or service.
- Know the percent state change value with the help of state transition
- Compare the value of this state change against low and high flapping thresholds
- When this value exceeds then the highest specified threshold then it is called flapping
- When this percent state value goes down the specified value then it is said that flapping has been stopped.
