Analyzing Flow Blockchain Transaction Failures

Dapper Labs has launched one of the most popular NFT gaming collections ever named CryptoKitties, which choked the Ethereum blockchain. There is no need to A NFT project called Cryptokitties crippled the Ethereum blockchain during the first major public cryptocurrency hype in 2018. Dapper Labs, a Canadian company, built the cryptokitties project and realized the popularity of cryptokitties, so they focused on scaling and improving their blockchain related projects. However, Ethereum did not go so smoothly since their NFT projects were competing for scarce resources with the DeFi dapps. In response, they created a new blockchain called Flow.

Flow is a fast, decentralized, and developer-friendly blockchain designed for high transaction throughput for gaming networks, digital assets, and mobile apps that require millions of transactions. 
Flow’s blockchain architecture follows a totally different route than Ethereum, with a multi-role architecture built on top of a unique architecture that handles scalability without sharding. 

The following chart shows the total number of transactions per per along with number of transactions failed as well as percentage of transactions failed. The first considerable spike in transactions failures were noticed on 13 May 2022 and it coincides with the sudden spike in new user registration. Though the number of new accounts being created on Flow trending high on the following days May 14th & May 15, the error rate has come down significantly.

The spike in errors on 13th May 2022 along with spike in account creations could be a correlation between them. Or the developers swiftly reacted to increased load and scaled out the ecosystem to to handle the sudden spike. We are not able to pinpoint the root cause why there was a spike in errors.

The error has been trending very high since May 25th 2022 and as I write this report on May 29th, the error rate is at whooping 75% for the smart contract “Flow Token” and “Flow Fees”.

When we look a bit deeper, we notice that most of the failure are caused by three types of events – Fee Deduction, Token Deposited and Token Withdrawn.

The following is shown on most of the failed transactions

[Error Code: 1101] cadence runtime error Execution failed: error: underflow

And it seems to be a code bug in one of the smart contracts according to flowjosh

you have an underflow somewhere, which means that you are trying to create a number that is below the minimum representation for the type you are using (UInt64). I would take a look at your arithmetic operations to make sure that is happening, or if you intend to have a negative number, use Int64 instead.

onflow forum

Leave a Reply