Foreword
Adrastia operates the described service on World Chain, having served around 931k individual price feed updates over the past month that Morpho consumes. For real-time and retrospective performance metrics and analytics, see the dashboard shared in the Live Deployments & Analytics section below.
Background
In DeFi, the reliability of pull-based oracle systems—where smart contracts must actively request external price updates—is critical to protocol safety. When price feeds are stale or inconsistently updated, lending markets face increased risk of bad debt, delayed liquidations, and inaccurate credit issuance. While pull-based oracle systems like Chainlink Data Streams and Pyth Core offer powerful infrastructure for high-quality price data, their security assumptions rely on the timely execution of updates. This makes the performance of feed updaters a key component of oracle resiliency. In this context, we introduce a managed solution designed to strengthen the operational layer of pull-based oracles through dedicated feed updater services, complete with real-time dashboards and live deployments across multiple networks.
Motivation
Pyth has demonstrated strong performance across many networks, reliably delivering high-quality pricing data with minimal latency. However, as with all pull-based oracle systems, the freshness of the data ultimately depends on external actors actively triggering updates. If a feed updater goes offline, even temporarily, the protocol relying on it may operate on stale data—posing serious risks in volatile markets.
Redundancy in feed update infrastructure is essential to mitigate this class of failure. Even the most robust oracle networks benefit from having multiple, independent operators contributing to data freshness. By diversifying the actors responsible for updates, protocols can reduce reliance on any single system and strengthen their overall oracle security model.
Chainlink Data Streams represents a promising evolution in pull-based oracles, offering high-frequency, low-latency data for protocols that demand precision. But like Pyth, it requires reliable off-chain execution to function as intended. Ensuring this execution layer is resilient, observable, and independently verifiable is critical as adoption grows.
This proposal introduces a service designed to address exactly that—adding operational redundancy and visibility to the feed update process for both Pyth and Chainlink Data Streams.
Solution Overview
To address the operational challenges of pull-based oracles, Adrastia has developed a robust updater system purpose-built for reliability, observability, and performance.
At the core is a custom updater client, designed to serve as an independent implementation—mirroring the benefits Ethereum enjoys from having multiple client designs. This diversity helps guard against systemic updater client bugs and adds an additional layer of operational assurance for protocols depending on timely price updates.
The infrastructure behind the client includes four bare metal servers, geographically distributed across multiple dedicated hosting providers. Each server is multi-homed to several RPC endpoints and, in the case of Pyth, to multiple Pythnet Hermes nodes—ensuring high connectivity and resistance to regional failures or RPC outages. This design creates multiple layers of redundancy and failover capacity in both network access and update execution.
In the unlikely event of degraded performance or failure, integrations with Datadog and Betterstack provide real-time observability and alerting, enabling rapid intervention. This monitoring layer covers both system health and on-chain update performance.
The system has already proven its reliability. On Gravity, our Pyth Price Feed Updater consistently achieves a median delay of less than 200ms from the moment a feed becomes stale to when a new update transaction is successfully broadcast to the network. This level of responsiveness enables protocols to trigger liquidations immediately upon market movement, which both maximizes profit opportunities for liquidators and minimizes the likelihood of bad debt accumulation.
Our Chainlink Data Streams Updater demonstrates equally strong performance. On World Chain mainnet, updates are consistently broadcast with a median delay of under 200ms from the moment a new report is available. This rapid response time ensures that the latest market data is made accessible to smart contracts with minimal delay, helping protocols react to market shifts with confidence.
Together, this system offers protocols a reliable, transparent, and performant solution for ensuring their oracle feeds remain fresh—whether using Pyth or Chainlink Data Streams.
Live Deployments & Analytics
Our updater infrastructure is actively deployed and operating on multiple production networks:
- World Chain mainnet – Chainlink Data Streams feeds
- Gravity mainnet – Pyth Core feeds
Both systems are backed by public real-time dashboards that provide full transparency into performance metrics, including update frequency, latency, transaction success rates, uptime, and more. These dashboards allow protocol teams and community members to independently verify that feeds are being updated on time and without disruption.
By exposing operational performance in real time, we enable protocols to confidently rely on our services without needing to blindly trust off-chain actors. This level of visibility is critical to building trust around oracle infrastructure and supporting governance-level decisions around oracle reliability.
We encourage everyone to check out our dashboards:
- World Chain mainnet – Chainlink Data Streams:
https://p.us5.datadoghq.com/sb/5977eb82-d451-11ee-8900-da7ad0900005-9c4d263e752acb0ed73963d0462fd945?fromUser=true&refresh_mode=sliding&tpl_var_batchid%5B0%5D=0-chainlink-data-streams&tpl_var_chain%5B0%5D=worldchain&tpl_var_customerid%5B0%5D=worldchain&tpl_var_environment%5B0%5D=production&tpl_var_groupid%5B0%5D=%2A&tpl_var_instance%5B0%5D=%2A&tpl_var_service%5B0%5D=adrastia-oracle-updater&tpl_var_unit%5B0%5D=adrastia-worldchain-0&from_ts=1747350370564&to_ts=1747955170564&live=true - Gravity mainnet – Pyth Core:
https://p.us5.datadoghq.com/sb/5977eb82-d451-11ee-8900-da7ad0900005-0db75e6110f6dcedd76d1fadb9987340?fromUser=true&refresh_mode=sliding&tpl_var_batchid%5B0%5D=0-pyth-feeds&tpl_var_chain%5B0%5D=gravity&tpl_var_customerid%5B0%5D=pyth-gravity&tpl_var_environment%5B0%5D=production&tpl_var_groupid%5B0%5D=%2A&from_ts=1745363125211&to_ts=1747955125211&live=true - Overall uptime:
https://status.adrastia.io/
Call to Action
If your pool relies on pull-based oracle systems such as Pyth Core or Chainlink Data Streams, we invite you to consider integrating with our updater services. Our infrastructure is built to deliver the reliability, transparency, and speed DeFi protocols require to operate safely under all market conditions.
To learn more, review our documentation:
- Chainlink Data Streams Feed Updater:
https://docs.adrastia.io/automatos/chainlink-data-streams-feed-updater - Pyth Price Feed Updater:
https://docs.adrastia.io/automatos/pyth-price-feed-updater
Feel free to DM me or reply here for questions or integration discussions.
Socials & Links
Website: https://adrastia.io/
Twitter: @AdrastiaOracle