One of the radio links in my network was down for 3 hours a few days ago. Why? An unintended interaction between parameters on the Ubiquiti radio on the clien

One of the radio links in my network was down for 3 hours a few days ago. Why? An unintended interaction between parameters on the Ubiquiti radio on the client side of the link collided with a DFS radar detection event on the access point side. The access point thought it had detected radar and randomly changed to a channel the client side wasn't scanning. As another ISP is using most of the 5GHz spectrum in the area, there are only a couple of channels that have low enough levels of interference to get decent performance out of the link. I had restricted the channels the client was scanning to reduce the recovery time from a loss of link a couple of years ago, but being groggy too early in the morning my brain took a trip to both ends to figure it out, and then more than an hour to find a channel that didn't suck. All of this is because 10 years ago when I talked to our Local Distribution Company that owns the poles on a 2km stretch where I want to run fibre to eliminate the radio link, the design tech said they would need to replace 10 poles at a cost of nearly $100k. Sorry, I'm not paying $100k to replace 10 poles that are near the end of their useful lifespan to place $750 worth of 24 count fibre. Complexity has unintended consequences: Lesson re-learned.