Understanding Connection Quality statistics (Application Connections)

Hi guys,

I am trying to understand where Pkt Loss registered by Connection Quality polling between two nodes and applications comes from. When reviewing the Node's packet loss statistics for the past 24 hours, all I see is 0 packet loss. This is on the Node level (polling IP).

I understand that this is packets loss registered between the nodes and NOT between the polling engine and the nodes. However, when I review the interface statistics for these two nodes, I don't find any packet loss registered there either.

So where is the packet loss occurring and where is it registered?

AFAIK, the Pkt Loss registered between the applications is inherited due to there being packet loss between the nodes. Is this correct?

ping

aLTeReGo

Find more posts tagged with

connection quality

packet loss

adm

Accepted answers

All comments

steffen.gutzeit

Heya,

Do you use agents instead of WMI/SNMP as your servers' polling method?

When you click on the connection it should show you details on the right hand side and you can go to the concerned connection for even more details. When you click that symbol you'll be redirected to the application dependency page: https://oriondemo.solarwinds.com/ui/sam.adm/linkdetailspage/src/140/dest/70

Best regards,

Steffen

Deltona

Hi there,

Both the nodes in my screenshot are being polled with WMI and have an agent installed for QoE and ADM purposes.

I have clicked on the symbol that you highlighted but I still fail to see why there is registered packet loss between the nodes/applications.

Where is the packet loss coming from, if not from the interfaces on the nodes?

jblankjblank

Deltona Packet Loss in an Application Dependency Connection is not based from the polling engine to the node. When applying the agent to those servers and choosing to enable "Connection Quality Polling" this is representing the quality of the connection between Server to Server, App to App, or App to Server. In your example you have both the servers and the apps on the map so you see two connections but in essence this is one, and having both sets of objects on the map may be redundant as the connection represented is identical.

"The Pkt Loss registered between the applications is inherited due to there being packet loss between the nodes." - This is correct.

Orion Maps is simply visualizing that info. The tool is in reality using Nping between the two endpoints.

serena may have further details.

Deltona

Thanks, I am now aware of where the packet loss is occurring.

What I need to know, however, is where from is the packets loss retrieved? I have looked at several network related counters on the nodes but they are all registering 0 in packet loss or errors.

serena

Deltona wrote:
Thanks, I am now aware of where the packet loss is occurring.
What I need to know, however, is where from is the packets loss retrieved? I have looked at several network related counters on the nodes but they are all registering 0 in packet loss or errors.

Hi!

Jeff is absolutely right, under the covers we are utilizing Nping - Network packet generation tool / ping utiliy to collect this packet loss information to measure the connection quality.