Spectrum ICMP device alarms only after 3 failed polls
Article ID: 272465
Updated On:
Products
Issue/Introduction
What is the best way to alarm on Pingable devices only after 3 or 2 failed polls?
We have a specific list of Pingable devices that we plan to monitor for UP/DOWN status. So if the device is DOWN, the default DOWN trap will be sent to Spectrum. If Spectrum polls and receives no respond, the default “DEVICE NOT RESPONDING TO POLLS (0x10009) ” alarm will be on the Spectrum. This is currently working for all SNMP and ICMP devices.
But now the request is not sending DOWN traps right away but it has to be DOWN more than 10 minutes only for devices in the list below. If I modify the DEVICE NOT RESPONDING TO POLLS (0x10009) cause code, it will effect all devices.
DX NetOps Spectrum ability to delay Alarm raise for down ICMP Pingable device for multiple consecutive poll cycles.
Environment
All supported DX NetOps Spectrum releases
Resolution
The request is an Enhancement Request (ER) that could be submitted to Product Management for consideration. Request an ER through Support via a new support case. Reference this Knowledge Base article.
Additional Information
Most alternatives revolve around global changes (Fault Isolation or DCM changes below), or ones difficult to administer long term for a specific or changing subset of devices (DCM changes). The global changes are ones that will impact all devices, not just the subset under review. They are not recommended.
- Fault Isolation ICMP Timeout and Try Count changes
- DCM Timeout and Retry changes.
- NOTE: Setting this on too many models could cause issues with model activation! This is not the best solution but it's possible.
- Changing timeout/retry values for the models under review following these recommendations.
- DCM Timeout to 60000
- DCM Retry to 10
- Poll Interval to 600
- Using Notifier to alert users via email?
- To only be notified (e-mailed) if the device is down for X time, create a SANM filter to age 10 minutes for these models.
Feedback
