Hello Matthew!
First, I want to thank you for reaching out! We love to engage with our community whether it’s answering questions, taking in suggestions or hearing success stories.
My first question is, are your sensors: wired, wireless, virtual/container or RWA (Remote Worker Agents)? That will help in understanding the necessary tweaking needed. I will go more into detail below, as for which tests to run, I am taking some idea’s I wrote about in our Targeting Best Practices.
What Should I Target?
While there is not a one size fits all approach that works for everyone, below are some suggestions based on extensive testing & feedback.
- Monitor Google or any extremely high uptime/availability resource
- VoIP applications
- Gateway test (especially for WiFi Agents/RWA)
- Your Cloud (You can also deploy an agent in the cloud and set it as a target too)
- Major SaaS applications used by your users
- Your datacenter
As for which type of test for each target, here are some guidelines:
- SaaS: All 4
- VoIP applications: Ping & PA
- Cloud: Ping, DNS & PA (HTTP if hosting SaaS Apps in your cloud)
- High Availability tests: All 4
- Datacenter: Ping & PA
False Positives
As for false positives, let me explain a bit on how our system tests to give some insight. This is why understanding what type of agents you have is important. Wired and virtual agents have a very high uptime and reliability, but this is only to the extend of your network. Wireless on the other hand is prone to more failures because the nature of wireless is far less reliable. Interference, channel saturation, physical/line of sight issues, etc. RWA (Remote worker agents), are even worse because you don’t have control over home networks at all.
For reference, the default ping configuration runs 17,280 times per agent, per test per day. So if you have 20 agents running 20 ping tests, that is just under 7 million tests per day. So almost any minor hiccups either in the network, across the internet or at the destination will be exposed. NetBeez will expose nearly every issue your network may encounter.
The question becomes, what is important, how quickly do you want to be notified of a problem? By default, five consecutive ping failures (running at a default interval of 5s) take 25 seconds to generate and an alert will be initiated. You can increase the time between ping tests (we recommend not more than 15s vs the default of 5s), or the number of required consecutive failures before the alert (we recommend 12 for noncritical applications).
Reporting
I wrote a post about reporting about 6 weeks ago titled NetBeez Reporting Best Practices so while I recommend you to check it out, I will give some feedback here as well. I would say a Network Status Report is the place to start. It will give availability and performance of each agent and in your case each branch as well as aggregated performance to each application. From there I would suggest generating a few target reports to key applications. That will break down site performance to those applications. You can also configure these to be sent to external personnel as long as they are either on the same network the server is hosted on or the server is publicly accessible, they do not need to have an account in the dashboard for this to work.
Conclusion
I hope this helps, I would be happy to expand further on anything but for now I don’t want to info dump too much. You can also reach me anytime at success@netbeez.net.