- The Engineering Workstream shipped 3 high priority features in December before the holiday:
- The swapper now supports the ability to trade from any ERC20 supported by Uniswap V3 to a Thorchain native asset adding thousands of new trading pairs that can be swapped in a single transaction via Thorchain.
- Added Multi-Hop / Multi-Transaction trades to the swapper via Lifi giving users potentially better pricing and access to more assets across EVMs
- Introduced Thorchain Lending support allowing users to access non liquidatable loans collateralized by BTC or ETH.
- The team got a much needed break over the holidays and spirits are high running into the new year as we focus on two primary features in January: Thorchain Liquidity Support and Memo-less swaps.
- Engineering and operations meeting this week to continually improve our release process and inter-team communication.
- Unchained and our self hosted infrastructure continue to be a source of down time for the app, breaking trust with our users and causing slowdowns to our release processes.
- 5+ production bugs were identified (and fixed) in the swapper during the month of December raising the need for us to invest more into the stability of the swapper as the primary driver of revenues from our application
December saw 2 large improvements shipped that greatly expanded the universe of assets available to swap within our application. Users now have the ability to swap between any ERC20 on uniswap v3 directly into a thorchain native asset in a single transaction. Additionally, we can now facilitate multi-hop trades in the UI, currently powered by LiFi, but built in a way to facilitate these types of transactions on any protocol in the future. Swapping in our UI continues to be a critical revenue stream for the DAO. While these additional trade routes are definitely an unlock to more trading pairs, December also regrettably exposed more than 5 production issues within the trade flow of the application. All of these issues have been resolved. Engineering, Operations, and Product have all aligned on treating production issues with swapping as the top priority in the coming months.
The team also deployed fixes to work around upstream issues regarding fees in the latest DOGE and BCH clients affecting sends and interaction with Thorchain Saver vaults. Multiple additional user experience improvements to Thorchain Saver Vaults were also rolled out to better handle edge cases and error messages to users when interacting with their vaults.
Our users were also impacted by 2 separate issues affecting the availability of our back-end infrastructure. Running our own self-hosted nodes, indexing layer, and APIs has continued to be a massive expense both in terms of infrastructure and hours spent on maintenance. Monthly AWS costs range between 20-30k USD per month and maintenance costs us roughly half of a full time engineer’s time. Solving the availability and costs problems is discussed more in the following sections.
Product and Engineering have aligned on two major feature releases in the coming months: full support for Thorchain Liquidity Providers and Memo-less swaps. Both of these features align directly with the “Thor + more” theme outlined by product and extend our support for the Thorchain ecosystem. Once released, users will have the ability to add liquidity, view their positions, and see summary statistics critical to their decision making. Support for memo-less will come in the form of a standalone application that allows users to trade using Thorchain by simply sending a very specific amount of assets to Thorchain designated addresses. The amount acts as a trade identifier and is tied to trade instructions via thorchain, allowing any wallet that can send assets to swap. Users of this new application will not be required to even connect a wallet in order to swap.
In addition to the exciting feature work the team will be focused on executing a series of tests with third party providers of nodes and indexing solutions to find a more cost effective and highly available solution. Previously, third party nodes have been utilized intermittently by the team, but due to the enormous request volume required by our indexing layer, these nodes were not cost effective nor performant. By utilizing a hosted solution for the indexing layer, we hope we can avoid this high call volume and effectively begin to farm out some of our data needs. It is reasonable to expect a hosted solution could cost a fraction of what we are currently paying, have much better SLAs for availability and free up precious engineering resources on our small team to focus on feature work.
First, I want to say thanks to the Engineering Workstream team and the community for the opportunity to step into this role. I am excited about the weeks ahead but also trying to remember that this role is temporary and temper the pace of changes to ensure the new leader has maximum flexibility when they take over. While I am leading the team, I do want to try to add additional transparency so that the community is more aware of our efforts and we can effectively share ideas and information beyond the core team. Working in a DAO has many challenges, but embracing the benefits and engaging the community to assist the core team will help us to realize the strength of the DAO model.
These monthly updates will be one small part in attempting to share information with the community about what we are working on, our successes and failures. I am open to any and all feedback from the community on what else might be helpful to ensure an open culture within engineering at the DAO.