I'd like to call out the apparent hidden service performance slowdown: https://metrics.torproject.org/torperf.html?start=2017-04-23&end=2018-01...
I hope the dev team is looking into it.
Thanks, Igor
On Sun, Jan 21, 2018 at 10:55:16AM -0800, Igor Mitrofanov wrote:
I'd like to call out the apparent hidden service performance slowdown: https://metrics.torproject.org/torperf.html?start=2017-04-23&end=2018-01...
I hope the dev team is looking into it.
That time frame matches exactly the overload issues we're dealing with.
See these previous posts for info: https://lists.torproject.org/pipermail/tor-relays/2017-December/014002.html https://lists.torproject.org/pipermail/tor-relays/2018-January/014175.html
That second post includes an explanation for why onion service performance will be more affected ("for one concrete example").
We're working on some mitigations that adventurous relay operators will be able to run to be more fair with their relay resources: https://bugs.torproject.org/24902 but they're still under review, so they're not quite at the stage yet where people should just grab them and apply them. :)
--Roger
On 22 Jan 2018, at 05:55, Igor Mitrofanov igor.n.mitrofanov@gmail.com wrote:
I'd like to call out the apparent hidden service performance slowdown: https://metrics.torproject.org/torperf.html?start=2017-04-23&end=2018-01...
I hope the dev team is looking into it.
Yes, we are: https://lists.torproject.org/pipermail/tor-relays/2018-January/014175.html
Because onion service circuits use more hops than exit circuits, they are affected more by network load.
T
tor-relays@lists.torproject.org