DFINITY Foundation

ICDevs.org

CryptoIsGood

Anvil

ICPL.app

ysyms

The Accumulators’ Neuron

RawTech Venture

Paul Young

ICPMANUAL

8YearGangDAO

Isaac Valadez

Krzysztof

Nicolas.ic

Smaug’s Neuron for Retail Investors

ICLight.io

Taggr Network

Sonic AMM

Rakeoff.io

OpenChat

ELNA AI

GEEKFACTORY

ICP Hub México

ICP Hub Poland

Jerry Banfield

WaterNeuron

ICP Hub Bulgaria

Gold DAO

NeuronPool

LORIMER ♾️ 🐶

B3Pay

ZenithCode

Aviate Labs

D-QUORUM

Yuku AI

CO.DELTA △

DragginCorp

Gwojda

Thyassa

🌮 TACO DAO Neuron

Ωmega-reject

Ωmega-vote

αlpha-vote

Cybersecurity

cyberowl

Dexter Caff/acc

JMVR

Smacker

HashEra

Install, reinstall or upgrade code of a canister controlled by the NNS.

# Proposal to upgrade the ckETH minter canister

Repository: `https://github.com/dfinity/ic.git`

Git hash: `3ce03862433d189f0d32c8a19cf99f502260ea98`

New compressed Wasm hash: `5a923e48361ae247b6e93a9daae16796ea3190e38658d4fd49776110889f4da0`

Upgrade args hash: `0fee102bd16b053022b69f2c65fd5e2f41d150ce9c214ac8731cfaf496ebda4e`

Target canister: `sv3dd-oaaaa-aaaar-qacoa-cai`

Previous ckETH minter proposal: https://dashboard.internetcomputer.org/proposal/144023

---

## Motivation

Bound and adapt how many transaction receipts the ckETH minter fetches per finalization round.

When two of the minter's four Ethereum RPC providers failed to reach consensus on `eth_getTransactionReceipt`, no withdrawal could be finalized, and the minter re-fetched the receipts of its entire pending set every few minutes. The outcalls grew from a few hundred a day to over 11,000 an hour, exhausting one provider's request quota and quadrupling the minter's cycle burn. The withdrawals themselves were executed on Ethereum throughout; only the minter's bookkeeping was stuck.

With this upgrade:
* The number of withdrawals whose receipts are fetched in a round is capped and adapts to the providers' behaviour: it doubles after a clean round, halves after a partly failed one and drops to one after a fully failed one.
* Receipts that were successfully retrieved are kept even if other lookups in the same round failed, so a backlog drains as providers recover.
* A withdrawal for which no receipt can be found no longer traps the minter; it stays pending and is retried later.
* New metrics expose the receipt fetch behaviour and the backlog of withdrawals and sweeps awaiting finalization.

Creating, signing and sending withdrawal transactions, as well as deposits, are not affected.


## Release Notes

```
git log --format='%C(auto) %h %s' 2c2c7c99bd526d8c3da51df32aeeb977e852292d..3ce03862433d189f0d32c8a19cf99f502260ea98 -- rs/ethereum/cketh/minter
3ce0386243 feat(cketh): DEFI-3013: Bound and adapt the per-round transaction receipt fan-out (#11636)
37681d65c8 feat(cketh): DEFI-3012: Expose the withdrawal finalization backlog as metrics (#11637)
 ```

## Upgrade args

```
git fetch
git checkout 3ce03862433d189f0d32c8a19cf99f502260ea98
didc encode '()' | xxd -r -p | sha256sum
```

## Wasm Verification

Verify that the hash of the gzipped WASM matches the proposed hash.

```
git fetch
git checkout 3ce03862433d189f0d32c8a19cf99f502260ea98
"./ci/container/build-ic.sh" "--canisters"
sha256sum ./artifacts/canisters/ic-cketh-minter.wasm.gz
```

Upgrade NNS Canister: sv3dd-oaaaa-aaaar-qacoa-cai to wasm with hash: 5a923e48361ae247b6e93a9daae16796ea3190e38658d4fd49776110889f4da0