This post is about the impact of network issues on services and applications. We will introduce all the relevant metrics like the number of retransmissions, etc. Of course, we will also show how to monitor those numbers easily and what that means for your environment.
Application Performance Monitoring is not just about measuring response times of services and applications. For being able to correlate and evaluate all of the collected data in a sophisticated way, processes and their network communication need also be taken into account.
With the latest release, our network monitoring support has reached full production quality and will be available to all customers updating their agents. By enabling this new, exciting functionality, Dynatrace will easily be able to answer the following questions:
Which processes consume most of my network resources?
Dynatrace always provided a meaningful overview of all the metrics that are important for easily gaining a first glance at network utilization. Now we also provide a comprehensive insight on the process level.
By highlighting a process, Dynatrace visualizes its contribution to the overall value inside the same chart.
Which processes experience network problems?
If a network link or segment is overloaded, it will start dropping packets. This is due to the TCP protocol, which tries to retransmit the packets that have not reached its destination due to congestion and packet drop in overloaded networks.
We detect such situations and visualize them as retransmissions. Due to Dynatrace correlation algorithms, we are able to do that on host and process level.
Connectivity of processes
Independent of physical networking issues like overloaded links, processes may have their own problems with utilization and stop accepting new network connections. From a client’s perspective, this results in a timeout or in a reset of an initial TCP handshake request.
Dynatrace tracks such issues as “TCP connection refused” and “TCP connection timeout”. As we not only provide data but answers, we compare that to the total number of connection attempts, which gives you a short and crisp overview of your processes’ network connectivity health.
Context is everything
Numbers and metrics are nice, having them available in every relevant context is even better.
Dynatrace not only displays those values on isolated detail pages, hidden beneath a plethora of other metrics. We work hard to stay true to our promise to not just deliver data, but answers. Therefore, the various problem contexts of Dynatrace always display everything that’s going on at the time the problem persists.
Network monitoring in Dynatrace really adds another level of value to your user experience. We provide all the data you need, where you need it, as Dynatrace sports a unified user interface, that allows you to gather all the necessary information without browsing endless plugin sub-pages. Above that, Dynatrace uses network domain problems as yet another potential aspect of its powerful root cause analysis. If the experience of your users is degraded due to network problems – you will know about this right away.
Update your agents and have a look at your network monitoring data. We are sure you’ll like the increased insight into the communication of your processes.