Router repeatedly turns on and off eth1

[NAVras]

I'm not sure what I did to cause this. I was trying a bunch of iptables and program commands on the router. Then I rebooted it after I'm done just to clear out any stray iptables command.
But then my PC can't connect to the LuCI, and I notice the Ethernet port is powered on and off repeatedly.
I logged into LuCI with another PC, and sure enough in system log I see the eth1 is being brought up and down repeatedly:

Fri Jul  9 20:30:19 2021 daemon.notice netifd: Network device 'eth1' link is up
Fri Jul  9 20:30:20 2021 kern.info kernel: [  701.510960] mt7530 mdio-bus:1f eth1: Link is Down
Fri Jul  9 20:30:20 2021 kern.info kernel: [  701.520618] br-lan: port 1(eth1) entered disabled state
Fri Jul  9 20:30:20 2021 daemon.notice netifd: Network device 'eth1' link is down
Fri Jul  9 20:30:22 2021 kern.info kernel: [  703.559456] mt7530 mdio-bus:1f eth1: Link is Up - 1Gbps/Full - flow control rx/tx
Fri Jul  9 20:30:22 2021 kern.info kernel: [  703.574627] br-lan: port 1(eth1) entered blocking state
Fri Jul  9 20:30:22 2021 kern.info kernel: [  703.585183] br-lan: port 1(eth1) entered forwarding state
Fri Jul  9 20:30:22 2021 daemon.notice netifd: Network device 'eth1' link is up

Any way to fix this?
OpenWrt 21.02rc3, router: UBNT ER-X

[NAVras]

Solved, phew...
I have the ip command to create custom table in startup script.
It was there for quite some time, but I guess it's a hidden bomb like suggested here:

Static IPv4 route not showing up in routing table

The standard method being unreliable possibly due to some sort of race condition that is usually not easy to troubleshoot, so try use options allowed_ips and route_allowed_ips in the WG-peer section.

I removed it and rebooted the router, all seems normal now.
I tried rebooting without removing them, didn't work...

Edit:
Use /etc/config/network for ip route and ip rule
For ip route there is also LuCI: Network --> Static Routes

[vgaetera]

To be fair, this doesn't look like a software issue.
The link up/down events are typically triggered by unstable L1 environment.

Also, I don't recommend using raw commands such as ip, iptables, etc.
Those aren't synchronized with netifd, so you're bound to run into race conditions.