Start small Define throughput/latency requirements BEFORE anything else, always let some room for scaling -> conclude on a default rate-limit Conduct proper smoke/load/stress tests using locust/k6 whatever Upgrade Hardware consequently Re-run the tests every day and after every upgrade to ensure quality is not impacted Questions for product: What's the expected usage of a client What are the most important endpoints

Pragmarrakech 2 Execution Feature: Implement Paradex/Hyperliquid Order Execution Specification: The system should be able to create, close, and modify positions on Hyperliquid and Paradex. Specification: The system should handle order confirmations, rejections, and potential errors from Paradex and Hyperliquid. Feature: Implement OMS & Tracking Specification: The system should consume orders from the Dispatcher.

Overview & Goal This document details the proposed strategy for publishing real-time financial market data, specifically prices and order book depth for both Spot and Perpetual markets, from the Pulse service into Apache Kafka. The primary goal is to establish a reliable, scalable, and well-defined data pipeline that can be easily consumed by downstream applications, starting with a new Rust SDK. This design prioritizes clarity, performance, and future extensibility. Context Data Producer: The Pulse service, which aggregates and processes data from various sources (Exchanges, Blockchains via Fetchers, Solvers, Pricers). Data Types: Price and Orderbook, qualified by Market Type (Spot/Perpetual). Message Broker: Apache Kafka. Consumers:

Post-mortem 10 Apr 2025 - EUR/USD This thursday at 4.30pm CET, there was an outage on Pragma's EUR/USD mark price feed. This document will outline the course of events that led to this outage and the next steps taken to ensure it doesn't happen for this feed and future feeds being integrated. Timeline In a discussion with our main data provider, we asked them to give us a test account so that we can test with new feeds without impacting the current API. They updated the credentials of our account without any notice which in turns resulted in an outage of our publishing service, as the connection got closed and login was failing. In the meantime, we used to have a secondary data provider as a fallback which also experienced some issue leading to incorrect decimals being published. This led to returning a mark price of 0.00011. This invalid mark price led to the liquidations of positions on Extended. Thanksfully all of the funds could be retrieved and all the users refunded.

Madara Committee What's the goal of this commitee ? The Madara committee is responsible for the growth and development of the Madara stack. It should thoroughly monitor the work of every actor involved. It should also be responsible for events, hackathons, communication around Madara. The committee is also in charge of setting up the best legal structure in order to recieve funding from the different actors of the ecosystem, and mitigate the risks of the contributors and companies building on top. What's the value accrual model towards Madara ? Madara is fully open-source, and the goal of the committee is not to make profits. The committee still should be able to get funds to fund development, and the most obvious way to do it is to hold the legal risk for the different companies using Madara, by running their sequencer(s) (cf: the Optimism foundation w/ the OP Stack). Other paths can be followed, with Gitcoin or grants from ecosystem, but it might be in a less scalable way. If Madara launches a shared sequencing layer, similar to the superchain, it's possible that we might have a token for economic security and fees. The token could also be used for governance in committee decisions.