Dear SolarWinds,
On case 01355411 Support confirmed for us that NPM does not support bgpPeerInUpdateElapsedTime OID, which present the correct uptime/change for BGP entries - established state.
Support explained to us that only following OIDs are supported and NPM is defining uptime time ("Last Change"), if SNMP polling discover a change in below values.
bgpPeerState 1.3.6.1.2.1.15.3.1.2
bgpPeerRemoteAddr 1.3.6.1.2.1.15.3.1.7
bgpPeerRemoteAs 1.3.6.1.2.1.15.3.1.9
I believe this is a bad design.
When polling BGP via SNMP (for example every 5 minutes), NPM could potentially miss the information about uptime/change (if BGP was not in the established state for 1-2 minutes).
This is exactly what we are seeing below, when comparing BGP “Last Change” to SSH console. Some BGP changes were visible during SNMP polling and some were not.
NPM is already doing a better approach with “Last Boot” datetime for Nodes, which provides very good knowledge of the correct node uptime.
In my opinion, NPM should do the same with BGP and bgpPeerInUpdateElapsedTime OID.
Please include support for bgpPeerInUpdateElapsedTime. You could also add new column for "Established for" datetime. For now, we will create a workaround with custom UnDP.
Thank you.
bgpPeerInUpdateElapsedTime
Description:
"Elapsed time (in seconds) since the last BGP UPDATE message was received from the peer. Each time bgpPeerInUpdates is incremented, the value of this object is set to zero (0)."
OID: 1.3.6.1.2.1.15.3.1.24
Comparison between SNPM output and "Routing Neighbors". Some "Last change" are correct, some are not.
Below example was from SolarWinds Support:
Thank you,
Marcin.
