Network request to Github is a dependency for Mycelium deployment

[scottyeager]

I saw this fatal error when trying to deploy to a Freefarm node on testnet:

error: could not deploy network net: error waiting deployment: 
workload net within deployment 56577 failed with error: 
failed to create network resource for network 504-56577-net: 
failed to setup mycelium: failed to get public mycelium peer list: 
Get "https://raw.githubusercontent.com/threefoldtech/zos-config/main/testing.json": 
context deadline exceeded (Client.Timeout exceeded while awaiting headers)

First of all it's pretty suspicious to me that our nodes are timing out trying to reach Github. I didn't see any evidence of an outage at Github at this time.

But the bigger issue is that it looks like we are potentially fetching this peers list file from Github during every deployment that includes Mycelium. I don't think the peers get updated frequently enough to justify this, and there's a big downside which is that it's another way deployments can fail.

Why don't we just fetch the peers every day/hour/whatever, cache them, and use the cached version for deployment?