Does a high API success rate mean every swap was completed?
When looking at a swap tool’s “success rate,” first check what it counts. A server successfully responding to a request is not the same as a swap being settled on-chain.
On October 28, Uniswap Labs launched a developer analytics tool that separates these two types of statistics: Success rate is the share of requests that received a successful response, with errors listed by API; Swaps separately tracks the number of swaps, their on-chain value, and average size.
So you can’t directly describe the first percentage as the “user swap completion rate.” Requests may come from different API categories, and the denominator does not automatically equal the number of attempted swap orders. Even if the API is healthy, you still need to check actual settlement records separately.
When researching data for $UNI -related applications, you can start by selecting the same time range, then distinguish between request categories, networks, and swap counts. If there’s only one overall success rate and its denominator isn’t clear, don’t use it to estimate how many users completed a transaction.
This article explains the statistical definitions in the announcement. It did not log in to any developer account or measure the current execution performance of any particular application. Clearer data categories can help troubleshoot issues, but they do not guarantee that specific transactions will settle.
Source: https://blog.uniswap.org/uniswap-developer-analytics-are-live
When looking at a swap tool’s “success rate,” first check what it counts. A server successfully responding to a request is not the same as a swap being settled on-chain.
On October 28, Uniswap Labs launched a developer analytics tool that separates these two types of statistics: Success rate is the share of requests that received a successful response, with errors listed by API; Swaps separately tracks the number of swaps, their on-chain value, and average size.
So you can’t directly describe the first percentage as the “user swap completion rate.” Requests may come from different API categories, and the denominator does not automatically equal the number of attempted swap orders. Even if the API is healthy, you still need to check actual settlement records separately.
When researching data for $UNI -related applications, you can start by selecting the same time range, then distinguish between request categories, networks, and swap counts. If there’s only one overall success rate and its denominator isn’t clear, don’t use it to estimate how many users completed a transaction.
This article explains the statistical definitions in the announcement. It did not log in to any developer account or measure the current execution performance of any particular application. Clearer data categories can help troubleshoot issues, but they do not guarantee that specific transactions will settle.
Source: https://blog.uniswap.org/uniswap-developer-analytics-are-live
