getting in ยท post 03 of 7 in this channel
Slow loading on a WTN onion address is the normal state
Pages load slowly here. Not sometimes, not when something is wrong. As a rule, and by design.
the three wethenorth market addresses
hn2paw7zfvndw3dovycegeqmvvnf4pl67b3g2p7pohjlzavloosh73id.onionhn2paw7zrgujyhnt6mgxlt2q6uhgbke4itpqitxhyfbumq3wtnckbuyd.onionhn2paw7zadwkcra3qzv5e4q547i7e5lvxm62cfxqftuqdu7moiu2ceyd.onionprinted the way they were handed over. no labels, no numbering, nothing here says which one to try first. this feed never opens them, so it has no status light, no percentage and no date of checking to give you.
where the delay comes from
A normal web request goes more or less straight to a server. A request to an onion service goes through a chain of relays, and so does the reply, and the two sides meet at a point neither of them chose.
That is a lot more hops than a direct connection, and every hop is a real machine somewhere with its own load and its own link.
The latency is not overhead that somebody forgot to optimise away. It is the arrangement doing what it was built to do.
what normal feels like
Several seconds before anything appears is ordinary. Ten or more is common. A page that hangs, then arrives all at once, is ordinary too.
Images and anything heavy will crawl compared to what you are used to. Markets built for this tend to keep pages light for exactly that reason.
None of that is a signal. It is the baseline, and the baseline moves around from one attempt to the next.
why misreading it matters
Somebody who reads slow as broken starts looking for a fix. The fixes on offer are the dangerous part: a different address from an unfamiliar page, a gateway that promises to be faster, a tool that claims to speed things up.
A slow page you waited out costs you thirty seconds. A replacement address you accepted out of impatience costs considerably more.
Impatience is the actual attack surface here, more than any technical thing.
the reload reflex
When a page stalls, most people hit reload. Here that is usually counterproductive, because it throws away the work in progress and starts the whole circuit business again.
Wait longer than feels reasonable before reloading. Then wait a bit more.
If you do reload and it stalls again, that still does not tell you much. It is one more sample from a noisy process.
when slowness is worth acting on
When it is total and persistent across all three addresses, over a decent stretch of time. That is a different shape from slow, and it is covered in a mirror that will not open.
Even then, the honest reading is limited. Persistent silence tells you your side is not getting through. It does not tell you why.
The useful distinction is between a page that is slow and a page that gives you nothing at all after a long wait. Those feel the same in the first twenty seconds and mean different things after two minutes.
Almost nobody waits two minutes. That is the gap where wrong conclusions get made, and it is worth knowing that you are standing in it.
replies
a reader askedIs one of the three WTN market mirrors faster than the others?
the answerThis feed does not open them, so it has no speed figures and will not invent any. Speed also varies by circuit and by moment, so a number measured elsewhere would not describe your connection anyway.
a reader askedCan I make Tor faster?
the answerNot meaningfully, and most things marketed as speeding it up work by removing the part that made it worth using. The security level setting does affect how heavy a page is, and that is covered in the security level.
a reader askedThe page loaded instantly. Is that suspicious?
the answerNot by itself. Circuits vary and cached pages are quick. Speed is a bad detector in both directions, which is the point this post keeps making.
what this post is not Not a benchmark. There are no load times measured here, because measuring them would mean opening the addresses, and this feed does not do that.