Hey, I run a Monad mainnet node and I spent a week remeasuring the FastLane auction from RelayFeeCollected / RelayBidFailed, then measured chain-wide priority fees off every receipt.
Window is 23 Aug 09:39 UTC to 30 Aug 09:39 UTC, exactly 7 days. Local archive only. No vendor indexer.
MEV Pulse is useful. I pulled a day of their GraphQL on 26 Aug and it sits in the same range as my counts. This is not a nobody-has-measured-this post. Pulse is MON-only and auction-only, and /methodology is still 404 (checked 6 Sep). Fine for a dashboard of one venue. Less fine if you want the rest of the chain.
Numbers below use MON at $0.02657 from Uni V3 WMON/USDC at the end of the window.
Auction bids collected: about $1.5k/day (58k MON).
Chain-wide priority fees: about $3.3k/day (125k MON).
Together: about $4.9k/day.
So the official auction is roughly a third of what people paid for ordering that week. Pulse shows the $1.5k. Two caveats: tips are not all MEV (normal users tip too), and I only measured what goes into the auction contract, not the validator / shMON split after.
From joining all 3.6M bid txs against the logs:
82% paid, 12% failed, 6% silent. Same shape as a 24h sample I ran earlier.
Every bid had exactly one target hash. Zero backrun bundles in about 2M blocks. The backrun channel in the contract is unused. That does not mean backrunning is impossible. People can still tip on normal txs, which matches the two thirds sitting outside the auction.
172 senders landed a winning bid, top 5 took 46% of fees. I did not cluster entities so 172 is an upper bound.
payBidOnFail was false on every submission. A losing bid only costs gas. The average loser declared about 40x what the average winner paid. Searchers declared 3.0M MON of bids and actually paid 407k MON. If you treat declared bid volume as demand, that number is cheap to inflate.
My own 24h sample on 23 Aug was about $2.9k/day on the auction. The 7-day mean is $1.5k. Peak to trough in the week is about 7x. I would not treat any single Pulse day, or mine, as typical.
No traces, so no searcher PnL. The gas figure I have is a ceiling (gasLimit x gasPrice), not gasUsed.
Writeup, method, scripts, JSON:
Happy to be corrected. I would rather catch a mistake here than see it copied.
Quang Nhan