cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post
Level 9

Using SNMP to determine device status

Jump to solution

Hi,

Does NPM use ping or SNMP to determine if a node is up? Is there an option to choose the number of consecutive missed pings before a node is set as down?

Tags (1)
0 Kudos
1 Solution
Level 14

wayneDigi,

NPM uses ICMP (ping) to determine up/down status.

While you can configure SNMP Retries on the Orion Polling Settings view (Click Settings in the top right of the web console, and then click Polling Settings in the Settings group), NPM does not count the number of ICMP retries for a missing object. Instead, if a device is unresponsive to ping, NPM initiates fast-polling mode, where the object in question is repeatedly sent ICMP packets. This variable fast-polling period is referred to as the Node Warning Level in the Calculations & Thresholds Orion Polling Settings view.

HTH,

EDIT: For more information about Orion Polling Settings, see "Orion Polling Settings" in the SolarWinds Orion NPM Administrator Guide.

View solution in original post

0 Kudos
8 Replies
Level 14

We are thinking disabling ICMP in the future for security purposes. So we will need this as well.

0 Kudos
Level 14

wayneDigi,

NPM uses ICMP (ping) to determine up/down status.

While you can configure SNMP Retries on the Orion Polling Settings view (Click Settings in the top right of the web console, and then click Polling Settings in the Settings group), NPM does not count the number of ICMP retries for a missing object. Instead, if a device is unresponsive to ping, NPM initiates fast-polling mode, where the object in question is repeatedly sent ICMP packets. This variable fast-polling period is referred to as the Node Warning Level in the Calculations & Thresholds Orion Polling Settings view.

HTH,

EDIT: For more information about Orion Polling Settings, see "Orion Polling Settings" in the SolarWinds Orion NPM Administrator Guide.

View solution in original post

0 Kudos

Thank you for your quick response!

Are there any immediate plans to use SNMP as well as ICMP to determine node status? Reason is, I have some Linux nodes that were down recently, but because ping was ok, the node appeared as up.

0 Kudos

wayneDigi,

I'll mark this thread for PM, so they can reply more authoritatively.

Thanks,

0 Kudos

Just my 2 cents.  But I vote it does not change from the way it is.  Ping is a very simple and reliable way to determine UP/DOWN, ICMP is very basic and reliable and part of the protocol stack where SNMP is technically in the application layer.

So if ping fails you know that on not even basic network functionality is there for one reason or another.  If SNMP fails then there are many more things that could cause that outside of just network functionality.

 

Again, just my 2 cents.

0 Kudos

Agreed, we test for SNMP failing separately as there can be a number of reasons why its failing (most common being a changed community string etc).

ICMP and Orion's aggressive polling during initial ping fail has been more than enough for us.

Andy.

0 Kudos

We are looking at adding this as an option the user can choose in the future, with ICMP as default.  There are use cases with some users where ICMP in blocked, but SNMP is allowed.

0 Kudos

Is there any movement on this request?  It looks like users have been asking about SNMP polling enhancement for years.

I agree that ICMP is the most basic and simplest method for determining node UP/DOWN and should continue as the primary method to determine UP/DOWN but there are still many times when ICMP is not being allowed.  My current situation is one where some devices are not able to be pinged.

I expect most other users have implemented another monitoring applications for these devices, which looks like what makes sense for me as well.

Please let me know where this request lies.

 

Thank you for your help,

Don

0 Kudos