Understanding Network Retransmission metrics

[Srikar_Mohan2]

Hello,

I am trying to understand what each of the "received (& sent) transmission" metrics mean. Unfortunately the description is not that helpful. For example: what does "base received" mean, what is the difference between the host aggregated process and the other retrans metrics, which metric is used under the "network overview" dashboard.

Also how are these metrics captured and at what frequency.

Thanks

Srikar

• builtin:host.net.packets.rxBaseReceived - Host aggregated process retransmission base received per second
• builtin:host.net.packets.rxReceived - Host aggregated process retransmitted packets received per
• builtin:tech.generic.network.packets.baseReRx - Retransmission base received per second on host
• builtin:tech.generic.network.packets.baseReRxAggr - Retransmission base received per second
• builtin:tech.generic.network.packets.reRx - Retransmitted packets received per second on host
• builtin:tech.generic.network.packets.reRxAggr - Retransmitted packets received per second

[AntonioSousa · Accepted solution]

I'll try to reply to some of your questions.

Retransmissions are a normal part of TCP, and sometimes even desirable (consider bufferbloat). Retransmissions can occur both in/out of a certain system, and most OSs have an indication of what is going on. "netstat -s" gives you results in Windows&Linux, so you can check it out. For more information, you should check out:
https://www.dynatrace.com/news/blog/detecting-network-errors-impact-on-services/

Of the metrics you refer to, #4 and #6 are at the process level, with the remaining being at the host level. There is really not a lot of information of what are the differences, but I would suggest you check the new metrics tab, where you can do a quick lookup of the metrics and their values.

The retransmissions in the Network Overview dashboard are a percentage, so none of the metrics you refer to. They are a composite value. You can check more information at the following link, which also gives more insight on the subject:
https://www.dynatrace.com/support/help/how-to-use-dynatrace/networks/how-to-monitor-network-communic...

Relating to capture, the first link discloses some of the ways that Dynatrace might use, and using system information will be certainly one of them. At the process level, I believe Dynatrace has not disclosed a lot, but Dynatrace does have network monitoring capacity, and it might also use some strategies that are possible at the operating system level. These metrics are typically available with a 1 minute resolution.