What is Osmosis?

First of all, Osmosis is an AMM (Automated Market Maker) which is built for and operates upon Cosmos. But it more than just an AMM since it provides the possibility to move seamlessly with low fees between IBC compatible block chains. To put it in other words, Osmosis has not fully accomplished it's aim that is connecting different dapps on different chains, but is not the path, currently focusing on cosmos chains. Also, you can say that Osmosis is decentralized exchange with proof of stake consensys algorithm which is built as a cross chain protocol from the foundation.

What are we going to see in this analysis?

Recently, there has been so many failed transactions on Osmosis. We want to see its stats, its possible reason(s) and a possible solution to the problem.

We are doing this for the past 4 months; as you can see in first chart there isn't any cases of high failed transactions before May 7th - the day that terra death spiral began. since Osmosis have many terra related pools with high TVL, it was expected that this create a huge impact wave on Osmosis.

There are 5 types of transactions types: TX, Coin-receinved, Coin-spent, transfer and message. There are some traits here; meaning a tx_id can contain various tx_type so that is the reason behind high number of failed TXs. In the search for a possible reason, I took a look at daily number of transaction; it is evident that there is direct relation between number of transactions and number of failed transaction. this suggests that there might be a congestion of network due to high tension dominating the market.

Out of those 5 types, TX had the most Failed count (it probably the most inclusive group) so I decided to break them all down to see what are these transaction? there are different names designated for each TX under the category of codespace; so the chart below shows how many of each type relate to which codespace, and it is obvious that gamm is responsible for most of these failed transactions.

So, one more break down: how does it look in weekly time period? As shown in chart below, starting from the May 9th Week, gamm codespace responsible for failed transactions grew significantly. SDK is also growing but there has been times in past that SDK was responsible with more failed transactions without a high number for gamm.

Conclusions

It can't be said with 100 percent certainty that what exactly is the reason behind these failed transactions, since on-cahin data provided on flipside tables are not enough. if there was information about pools for example, we could see how these transactions distribute amongst pools and that could make a big difference. nevertheless, I mainly can speak of two reason for it:

• Congestion of network on a high tension market, where people easily attempt to move their assets (low fee), trading for SC, etc. I'm not sure how nodes are operating in Osmosis but any attempt to increase scalability would prevent these situations.
• Slippage; if too many people start to withdraw their liquidity, many transactions will fail since slippage is changing dramatically in every second.