Transaction Failure
Establish the failed transaction rate for Terra over a time period of your choosing. Is the referenced tweet correct?
In this dashboard, I gathered the Rate of Failed Transactions since October 2020. I compared my result with the referenced tweet that you can see below, reviewed some reasons in my opinion, and provided a few recommendations to prevent failed transactions.
In the following table and graphs, you can see all the numbers of "Total Transactions", "Successful Transactions", "Failed Transactions" and "Failed Rate".
Note: For having Fair results I decided to not consider March 2022 in my data(It is not a full month yet)
As you can see my result and referenced tweet are almost the same which shows that the tweet is accurate.
But In my opinion, having ((A Large number of transactions)) is the main reason for increases in the failed transactions and one of the events that we have lots of transactions in a short amount of time is NFT minting.
By looking at the Overall Graph, you can see that since October we had an increase in the number of failed transactions also October was the month that we had the first NFT minting on Terra Network( Galacti Punks)
According to their website, all Galactic Punks were minted on October 2nd and sold out in 8 minutes. If I remember correctly on that date my transaction for minting an NFT failed at least 5 times which shows the minting process with lots of users applies heavy pressure on the network.
In my opinion, there are two reasons for having failed transactions:
1- Problems in the Network
2- A large number of transactions
In the past few months, we didn't have major incidents on the Terra network. According to Terra Status we had 1 major incident in the last 6 months which was ((Bombay public node outage)).
For the first reason (Problems in the Network), we can do many things to improve the network which costs the system a lot, but for the second reason (A large number of transactions) and in my opinion, the most important reason we can do many improvements without spending money.
The first improvement is applying whitelists. By doing that we decrease the number of users and subsequently fewer transactions, but again as you can see it is not enough. Projects should apply different levels or groups for their whitelists. for example, instead of having 10000 users in 1 whitelist and 1 minting period, they can divide them into 5 groups (each group with 2000) and their own minting period. By doing that we will decrease the pressure on the network and probably we will have fewer failed transactions.