Prysm beacon client can't hold outbound peers for more than a few hours

[honcho26]

Describe the bug

I am running Prysm v5.1.2 as a beacon client with a Geth 1.14.12 execution client, and a validation client using Prysm v.5.1.2 as well. When they start up, everything works great at first. The node synchs and starts validating. However, within a few hours, it drops down to zero outbound clients and starts missing attestations. Inbound peer numbers remain satisfactory. As the outbound peers fall off, I also start to see a few attestations with non-zero optimal inclusion distances.

Here is the prysm command line I'm using for the beacon client, with personal info/addresses obscured:

G:\prysm\prysm.bat beacon-chain --datadir (path) --execution-endpoint=http://localhost:8600 --jwt-secret=(path) --suggested-fee-recipient=(walletaddress) --checkpoint-sync-url=https://beaconstate.info --genesis-beacon-api-url=https://beaconstate.info --p2p-host-ip=(ipAddress)

I have opened all of the recommended TCP and UDP ports and forwarded them to my validator machine through my router.

Has this worked before in a previous version?

Yes, this has only been happening for the past 6 months or so, but has recently gotten worse.

🔬 Minimal Reproduction

No response

Error

No response

Platform(s)

Windows (x86)

What version of Prysm are you running? (Which release)

v5.1.2

Anything else relevant (validator index / public key)?

Validator index 1408548

[nisdas]

In our next release which should be pretty soon can you can run with --enable-discovery-reboot ? It was added here for windows users specifically: #14487