[tor-talk] Metrics user-count beta related to performance graph?

[Sebastian G. <bastik.tor>]

Hi,

(Probably only Karsten knows).

The new estimated directly connecting users graph (beta) [1] shows a
little slow-down in growth, a little spike and a little decline in
counted users.

The performance graph for 50KiB pieces [2] shows also a decline in
download time, which seems to be starting before 0.2.4.17-rc.

Does it mean that your new approach in counting users is more accurate?
(or is it too early to say?) (or are Tor related changes to be accounted
for the better performance)

(I still expect the botnet to hit its saturation point, where every
infected machine uses Tor. Then seeing spikes [or not if too tiny] for
new infections and some decline for disinfected machines.)

[1] https://metrics.torproject.org/users.html#userstats
[2] https://metrics.torproject.org/performance.html#torperf

Sebastian G.