Well, the manual does describe this technically, but it doesn't provide much, if anything for examples, which would be helpful.

It seems to allow you to add the values from various sensors. For example I wanted to get the aggregate total of 3 sensor input and output.

What I used was:

1. 1:XO-in
   Channel(2123,0)+Channel(2103,0)+Channel(2163,0)
   
2. 2:XO-out
   Channel(2123,1)+Channel(2103,1)+Channel(2163,1)
   
3. 3
   100000

These are collected at the remote routers where they all come in to the data center in a single connection. This connection is smaller than the sum of the allowed traffic of these three, so we need to know if we hit the ceiling which is 100meg.

One of my favorite uses for Sensor Factory Sensors is for notifying me of network connectivity problems. While all Sensors do this, I use the Sensor Factory to eliminate false positives.

Here's what I do. I have 3 remote probes, all with the same sensors. These remote probes are located around the country running on different cloud providers.

So lets say I have a ping sensor on each probe pinging my Data Center's router's WAN side IP. The Sensor Factory "rolls these up" in one graph.

This is good, because although I may see a spike in latency from one probe, if I don't see it on all 3, I know it was an external issue. before this, it was unclear whether there the problem was on my end or not and required further investigation. (wasting my time)

Now for notifications, I shoot out text messages/email to our ops group if the sensor is down. (this requires all 3 remote probes to show down) Having this 'majority vote' has effectively eliminated false positives related to external network problems and has made anyone on our CriticalAlert list very pleased that if there is a Notification, we know it is real and serious.

Saves us from the 'boy who cried wold syndrome'

Just my 2 cents. but is working great for us.

-Ryan

Ryan,

Can you tell me how you got the probes in each data-center to communicate to allow the sensor factory to access their channels?

Thanks, James

Dear James

The probes all send their data to the server. The factory sensor runs on the core, even if it is put on a probe. That is why a factory sensor can access other probes.

Arne,

Thank you for your reply. I realize they all send their data to the core however if I had thee servers running in different cloud providers as in Ryan's example they would all need some form of connectivity between each other in order for the core/ sensor factory to access their channels.

In my scenario we have our kit hosted in one DC which is broken up into VLANS, we have a core server and probe servers monitoring sensors on each. what I wanted was one of the servers inside the DC to monitor for example the ISP public link and also one completely separate outside in say another DC monitoring it also. I would then aggregate these two sensors into a sensor factory sensor and only show down status if both links are down.

I just wasn't sure how to get PRTG to get data from the external probe in order to access the channel.

I hope that makes sense.

Thanks, James.

Dear James

Are you still speaking of getting data from a probe, or getting data from a core server which is not connected to the server which runs the factory sensor? It is not possible to get data from a separate PRTG installation using the factory.

Arne,

I am talking about getting data from the probe. So I would have a core server inside our DC (core where the sensor factory would also run) which pings the ISP IP and a probe running on an external cloud provider doing the same. is there a proxy inside PRTG or some such thing that would let me connect to the probe from the server to access the channel data?

Thanks, James.

Dear James

If the probe is connected to the PRTG core server, this implies that the core continuously get its data. Otherwise, you cannot get its data. The PRTG probe connection supports no proxy server. Any PRTG probe can connect to one core server only.

I notice in the Knowledge Base that the sample Factory Sensor has "N/A" for both Downtime and Uptime, as seen below. Is it not able to keep track of it's downtime and uptime?

Dear Evan

The factory sensor also allows you to change the formula at any point in time and has no historic data of its own, there is no track of downtime either.