Keep in mind that Chrome is not on the EPIC, so that test is not really of any value other than saying the gateway and network at the time of the test to the PC is fine.
This is why it needs to be tested from the EPIC and the groov Manage tools are a quick sanity check.
Since you are saying that sometimes it works, the next thing I would be quickly setting up is a Node-RED to groov View ping check running 24*7.
Install the node-red-contrib-advanced-ping in the pallete.
It outputs a millisecond ping time vs the regular ping node that just outputs true/false.
Take the output and put it into a groov View data store on the EPIC (so it is all on localhost - this is important!!).
Then trend it.
Here is what it should look like:
Of course, you will only have one trend line.
Ping the exact server of the URL you are trying to hit. (The server, not the URL).
I’d make a few tends of different lengths like I did in that post.
15 minute, and 1 week at the very least.
Now you have a long running trend from the EPIC and will be able to see when its up and when its down.
There is clearly something going on with the network connectivity from the EPIC to the server that is hard to spot / troubleshoot.