ICMP Polling

ICMP polling is a network monitoring technique that uses ICMP (Internet Control Message Protocol) packets to check device availability and measure response times. The system uses the fastest response time from the packet batch for monitoring.

ThirdEye’s ICMP monitoring system provides essential network availability tracking through optimized ping-based checks. It features configurable polling intervals (30s-5min,) and adaptive retry logic that automatically performs up to additional checks with dynamically calculated intervals when packet loss occurs. Key characteristics include:

• Dual-Packet Metrics: Sends 1-2 ICMP packets per cycle, recording the fastest response time
• Smart Retry System: Auto-retries with geometrically increasing intervals during failures
• Zero-Config Defaults: Auto-assigns “ICMP Ping (Default)” monitor to new devices with immediate activation
• Conditional Alerting: Supports thresholds for consecutive failures and RTT limits

ThirdEye’s ICMP monitor consists of the following settings:

• Interval
• ICMP Send Count
• Retry

ICMP timeout is always 2 seconds and cannot be changed.

Operation image 1

Setting details

Explanation

If you set the interval to 30 seconds, a ping (in this case 1 time) and 5 retries will be executed within 30 seconds. The retry interval is dynamically averaged based on the monitor’s polling interval, here 5.2 seconds.

Operation image 2

Setting details

Explanation

If the ICMP transmission count is “send 2 times”, pings will be sent 2 times and then retries will be performed 5 times.

The retry interval is dynamically averaged based on the monitor’s polling interval, but is up to 25 seconds, so a long interval will perform as shown above.

Time required until alert occurs:

Theoretical value: 30 seconds (2+5.2*5+2) if the interval is set to 30 seconds.

Additionally, ThirdEye has “response confirmation” and “period” as triggers for generating alerts.

In the response confirmation trigger, you can use “count” and “period” to generate an alert if “there is no response N times within a certain period of time.”

In the below case, an alert will be generated if there is no response twice within 3 minutes.

Sample image

In period triggers, you can use “conditions” in addition to “count” and “period” of response confirmation triggers. The “condition” can be the round trip time (RTT) of the ping response packet and the packet loss percentage.

By using these conditions together, it is possible to perform monitoring. For example, even if a ping response is returned from the monitoring target, the RTT does not reach the level expected by the user, so it is judged as NG and an alert is generated.

Sample image