What is their purpose?
All NAM monitors have an obligatory step-level constraint. It is set on the Success rate percent and indicates the condition that needs to be fulfilled to consider the whole step as passed.
Success rate percent is defined as the ratio of the number of requests that didn’t fail to all requests. For example, if your NAM monitor contains 6 requests and 1 request out of 6 has failed, the ratio is (6-1)/6 = 83.33%. If your constraint is defined as ‘>80%’, such execution of the monitor will be considered successful, as 83,3% is above 80%. But if you change your step-level constraint to 100%, the same scenario will cause the monitor’s failure.
Request-level constraints also define conditions that need to be met to consider each request successful. Request-level constraints are different for different types of NAM monitors. Let’s focus on ICMP requests, as this is the most popular type of NAM monitor.
We have an ICMP success rate percent constraint, defined as the ratio of the number of packets received to the number of packets sent. For example, if 5 packets (this is the configuration element) were sent and 4 were received, the ratio is 4/5 = 80.00%. If you define your ICMP success rate constraint as ‘>75%’, such request is considered successful, but if you define your ICMP success rate constraint as ‘=100%’, such request is considered as failed.
Have a look at the diagram illustrating how step-level constraints and request-level constraints work together for NAM ICMP Ping monitor.
HI @Jacek_Janowicz ,
Thanks for the Insights,
We’re using NAM ICMP synthetic monitors to monitor the availability of 400+ network devices. Each device IP address has its own synthetic monitor, and each monitor sends four ICMP packets per execution.
We want to distinguish two conditions:
- Unavailable: The device returns 0 of 4 replies for three consecutive executions. This should alert our network team.
- Degraded reachability: The device returns only 1, 2, or 3 of 4 replies. We want to see the packet loss and, if possible, raise a separate degradation alert without marking the device unavailable.
Apart from the Anomaly Detector that we can create. What thresholds we can use that will fit this scenario.
We currently have ICMP Success rate ≥25% and Success rate ≥25%.
For a monitor with one IP address, should the step-level Success rate constraint be ≥100%?
Thank you for your support in advance.
Regards
