Sounds like a kde issue

I took a quick look yesterday and saw some similar reports of flatpaks on kde requiring workarounds to have an icon. Didnt read too much into it though

<b​igmenpixel:matrix.org> I would be great but not necessarily

<r​ucknium:monero.social> According to my preliminary analysis, the share of non-spam txs using the 3rd fee tier (320 nanoneros per byte) increased from 4.5% on March 11 to 35.3% on March 12. So far, March 13 looks the same (35.6%).

<r​ucknium:monero.social> I thought the fix was supposed to bump auto fees from the 1st tier to the 2nd. But the PR says it will bump to priority `2`, which is 3rd tier because it looks like it is indexed from 0.

<r​ucknium:monero.social> Anyway, miners are probably happy. The first tier is 20 nanoneros per byte. Going to the 3rd tier raises fee per tx by 10x.

rucknium: it reduces the priority by 1 later in the code

1 = priority 0, 2 = priority 1

<r​ucknium:monero.social> This is why I try to avoid reading wallet2.cpp directly :P

but i'm confused now

why did the third tier increase so much?

<r​ucknium:monero.social> So PR 9220 might not have caused the change in the data I see

<r​ucknium:monero.social> I don't know. Could be a service, another wallet implementation, etc

<r​ucknium:monero.social> I'm also trying to code quickly and I haven't double checked everything. Could be a problem in my analysis code.

320 nXMR per byte is not one of the auto levels

Auto chooses between 20 and 80

<r​ucknium:monero.social> sech1: Are you seeing the same thing in the fee data? I'm just doubting my numbers. It looks like mean fee about doubled on March 12: bitinfocharts.com/comparison/monero-transactionfees.html#3m

If some users started using 80 instead of 20, it checks out

<r​ucknium:monero.social> This block has 121 txs paying 320 nanoneros/byte according to my count: xmrchain.net/block/3103997

<r​ucknium:monero.social> I think this could be a change in the spam behavior.

i just confirmed auto-fee setting the correct tier

jumps between first and second tier, around 4x difference

<r​ucknium:monero.social> Thanks, selsta. Fee tiers by blocks today: paste.debian.net/hidden/cedfac12

i don't see why the spammer would use the third tier

<r​ucknium:monero.social> It looks like spam behavior. 320 nanoneros/byte is high until block 3104043. That's why I was doubting my data: I didn't see a lot of these 320 fee txs in the very recent blocks.

<r​ucknium:monero.social> Not necessarily a spammer, but someone who produces a lot of txs on chain and can switch fee tier at will.

<r​ucknium:monero.social> Maybe Kraken or someone like that.

<r​ucknium:monero.social> xmrchain.net/search?value=3103829 a block of 100% txs with 320+ fee. Total fees in block 0.146910 wow

<r​ucknium:monero.social> It's those 1in/2out txs again, but at a higher fee.

if their goal is to create outputs to deanon it wouldn't make sense to pay 10x fees

<r​ucknium:monero.social> Well, you can own an even higher share of outputs if your txs displace real user txs in a block.

Real txs get mined eventually so it doesn't matter in the long run

<r​ucknium:monero.social> I saw a sudden increase of 320 nanonero/byte tx on March 7. The share of those txs was 13.5% on that day. I didn't know what it was. Now it looks related to what we are seeing yesterday and today.

recent blocks I'm seeing >90% min fee txs

just looked at full blocks, not all of them are currently full

<r​ucknium:monero.social> The 320 fee volume stopped at block 3104043