While I was checking fallback directory mirrors for #17158, I encountered some relays that took more than a minute to serve a consensus. Most took 150 seconds, which could be caused by a RelayBandwidthRate of 10 kilobytes a second.
One solution to this issue is to disable the DirPort on relays with a RelayBandwidthRate less than 50 kilobytes a second, or more than 30s to serve a consensus.
(This ticket also contains the python / stem code that I'm using to check the consensus download speed.)
But I don't actually know if it's the relays' torrc configuration that's causing the issue, or if their provider is limiting their speeds, or if there's some other issue.
If you're the operator of one of the relays listed below, can you let me know?
And if someone wants to investigate further … that would be great.