# retry Published articles for retry. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## The User Write Lock: an Alternative to Transactions for Google App Engine DevFeed: [The User Write Lock: an Alternative to Transactions for Google App Engine]() Original publisher: [Read original article]() Author: Khan Academy Published: 2016-06-20T22:00:00Z Content type: tutorial Language: en Sources: [Khan Academy]() Topics: [Transactions](), [Databases](), [Google]() Tags: [caching](), [databases](), [engineering](), [google](), [infrastructure](), [news](), [processes](), [retry](), [transactions]() ### AI overview This developer article explains how transactions in Google App Engine's datastore provide atomicity for single-entity and multi-entity writes, but can introduce frozen datastore state, retries, idempotence requirements, and restrictions on datastore queries. It presents the User Write Lock as an alternative, although the supplied text ends before describing that alternative. ### Source excerpt By Craig Silverstein Transactions are the standard method in Google App Engine's datastore -- as in many databases ... Read more ## TypeSafe AI's Jev now available on AI Gateway DevFeed: [TypeSafe AI's Jev now available on AI Gateway]() Original publisher: [Read original article]() Author: Jerilyn Zheng Published: 2026-09-16T00:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [gateway](), [SDKs](), [API](), [Software]() Tags: [ai-gateway](), [api](), [evaluation](), [gateway](), [model](), [queue](), [retry](), [sdk](), [software](), [support](), [typesafe-ai]() ### AI overview Vercel's AI Gateway now supports TypeSafe AI's Jev, a probabilistic decision model that returns typed Choice, Score, and Boolean answers with probabilities. The changelog describes parallel evaluation, AI SDK 7 integration through an experimental evaluate API, example routing and review workflows, and provider options including Zero Data Retention and No Training. ### Source excerpt Jev from TypeSafe AI is now available on AI Gateway. Jev is a probabilistic decision model for software: state goes in, typed Choice, Score, and Boolean answers come out. Regular language models generate text one token at a time, which the application then parses and validates. Jev evaluates all declared questions in parallel and returns typed answers plus probabilities directly. That removes unnecessary text generation and makes it straightforward to automate clear cases while routing uncertain ones to review. TypeSafe reports Jev was up to 193.6x faster and 444.6x cheaper than LLMs on its workflow evaluations. Example use cases include: Choosing the next tool or subagent in an agent loop Deciding whether to continue, retry, ask the user, or stop Scoring urgency or risk before an action Verifying model outputs and enforcing guardrails. AI SDK 7 exposes Jev through the experimental evaluate API. Choice selects an option, Score grades an ordered rubric, and Boolean estimates the probability of true. Install the current AI SDK (AI SDK 7.0.105 onwards supports the evaluate API): Each evaluation specifies: model: the evaluation model to call, state: the shared string, object, or array to evaluate, and questions: a map of named decisions to make about that state. Call the model with typesafe-ai/jev. This example turns one support case into a queue, priority, and refund-review decision, with uncertain routing sent for manual review: The result preserves question IDs and Choice keys. TypeSafe reports separate Choice and Score confidence in result.providerMetadata.typesafe.confidence. Calibrate probabilities and confidence against labeled examples from your workflow. Jev supports Zero Data Retention and No Training, enabled per request in the example. Evaluation calls also appear in logs and custom reporting, count toward budgets, and accept other Gateway provider options in the same providerOptions.gateway object. Read the documentation on evaluation models on AI Gateway f ## AI agent approval workflows: handling edits and retries DevFeed: [AI agent approval workflows: handling edits and retries]() Original publisher: [Read original article]() Author: WorkOS Published: 2026-09-15T15:02:39Z Content type: tutorial Language: en Sources: [WorkOS Blog]() Topics: [AI Agent](), [Authorization](), [Tool]() Tags: [agent](), [authorization](), [execution](), [how-to](), [retry](), [review](), [server](), [tool](), [workflows]() ### AI overview This guide explains how to add human approval to AI agent tool calls with WorkOS Airlock. It covers binding approval to an exact request, persisting the proposed action, resuming execution after approval, and recovering when a provider response is uncertain. ### Source excerpt Add human approval to AI agent tool calls. Learn how to review exact requests, resume execution, and handle edits and retries with Airlock. ## AI safety does not stop at the model DevFeed: [AI safety does not stop at the model]() Original publisher: [Read original article]() Author: Samar Abbas Published: 2026-09-15T00:00:00Z Content type: opinion Language: en Sources: [Temporal Blog]() Topics: [ai safety](), [Agent Harness](), [reliability](), [retry](), [Security]() Tags: [agent](), [agent-harness](), [ai](), [reliability](), [retry](), [safety](), [security](), [temporal-voices]() ### AI overview The article argues that AI safety extends beyond model behavior to the application layer, where companies must control agent authority and enforce approvals, policies, and credential limits. It emphasizes that these controls must remain reliable through crashes, timeouts, and retries, and presents Temporal Agent Harness as an execution-layer control point around an agent SDK. ### Source excerpt Auditing what an agent did is only half the job. Companies also have to control what agents may do, and make those limits hold when systems fail. ## Patching This Site's MCP Endpoint and Limiting What It Can Do DevFeed: [Patching This Site's MCP Endpoint and Limiting What It Can Do]() Original publisher: [Read original article]() Author: Brent Leekley Published: 2026-09-08T00:00:00Z Content type: tutorial Language: en Sources: [brent.leekley.me blog]() Topics: [Model Context Protocol](), [MCP Server](), [API](), [HTTP](), [JSON]() Tags: [aeo](), [agent-ready-website](), [agentic-ai](), [agents](), [api](), [api-security](), [clients](), [json-rpc](), [logs](), [mcp](), [mcp-server](), [model-context-protocol](), [prompt-injection](), [protocol](), [rate-limiting](), [retry](), [server](), [streamable-http](), [webmcp]() ### AI overview This article explains how to patch a website's public, unauthenticated MCP endpoint. It covers updating to the current protocol revision while retaining compatibility with older clients, and applying limits including read-only tools, per-IP rate limiting, request-size caps, Origin checks, POST-only access, non-reflected strings, and privacy-preserving logs. ### Source excerpt An MCP server on your website is a public, unauthenticated API. Bringing one to the current protocol revision while keeping older clients working, and the seven limits that bound it: read-only tools, a per-IP rate limit that always sends Retry-After, hard caps on body and argument size, an https-only Origin check, POST only, no reflected strings, and logs that store a salted hash instead of an address. ## How to Deal With Errors and Failures in LLM-Powered Applications DevFeed: [How to Deal With Errors and Failures in LLM-Powered Applications]() Original publisher: [Read original article]() Author: ByteByteGo Published: 2026-09-07T15:31:35Z Content type: tutorial Language: en Sources: [ByteByteGo]() Topics: [Large Language Model](), [App](), [Language models]() Tags: [applications](), [concurrency-control](), [errors](), [how-to](), [llm](), [rate-limiting](), [resiliency](), [retry]() ### AI overview A tutorial on error handling and resiliency for LLM-powered applications. It covers failures such as unavailable networks, rejected requests, invalid JSON, hallucinations, slow responses, timeouts, retries, fallbacks, circuit breakers, rate limiting, queues, concurrency control, safe tool calls, and response streaming. ### Source excerpt Apart from normal processing, the application also sends data to a large language model (LLM). It then uses the model's response to carry out a task. ## When the action succeeds and the audit write doesn't DevFeed: [When the action succeeds and the audit write doesn't]() Original publisher: [Read original article]() Author: WorkOS Published: 2026-09-04T00:00:00Z Content type: article Language: en Sources: [WorkOS Blog]() Topics: [Logging](), [systems](), [payload]() Tags: [audit](), [distributed-transaction](), [logging](), [metric](), [retry]() ### AI overview The article examines why audit logs can become unreliable when an application performs an action and writes its audit record as separate operations. It describes silent gaps, phantom events, and duplicate events caused by failures, ordering, and retries, emphasizing that audit records must remain trustworthy evidence. ### Source excerpt When the action succeeds and the log write fails, you get a gap that nothing alerts on. Here is the write path that closes it. ## Orchestrating Playwright test suites with Temporal for independent retries DevFeed: [Orchestrating Playwright test suites with Temporal for independent retries]() Original publisher: [Read original article]() Author: Simon Emms Published: 2026-09-03T00:00:00Z Content type: article Language: en Sources: [Temporal Blog]() Topics: [Playwright](), [Orchestration](), [ci](), [test]() Tags: [browsers](), [ci](), [orchestration](), [playwright](), [retry](), [temporal-voices](), [tests]() ### AI overview This article explores using Temporal to orchestrate Playwright end-to-end test suites as independent, durable units of work. The approach allows failed tests to retry independently while preserving progress instead of restarting the entire CI job. ### Source excerpt Learn how Temporal can orchestrate end-to-end test suites so failed tests retry independently, preserving progress without restarting the entire CI job. ## Refresh token behavior across fourteen providers DevFeed: [Refresh token behavior across fourteen providers]() Original publisher: [Read original article]() Author: WorkOS Published: 2026-08-31T00:00:00Z Content type: article Language: en Sources: [WorkOS Blog]() Topics: [OAuth](), [API]() Tags: [providers](), [retry](), [rotation](), [token](), [tokens]() ### AI overview This article compares documented OAuth refresh-token behavior across fourteen providers, including token rotation, expiration reporting, grace periods, and revocation on reuse. It explains how rotation affects persistence and retry handling, with examples including Google, Box, Atlassian, Dropbox, Asana, GitHub, and Salesforce. ### Source excerpt Which providers rotate refresh tokens, which return expires_in, which give you a grace period, and which revoke on reuse. One row per provider, verified against provider documentation in August 2026. ## 【kube-apiserver】运维与升级：HA、flags、graceful shutdown 与 etcd 联检 DevFeed: [【kube-apiserver】运维与升级：HA、flags、graceful shutdown 与 etcd 联检]() Original publisher: [Read original article]() Author: Liao Tonglang Published: 2026-08-28T00:00:00Z Content type: tutorial Language: zh Sources: [土法炼钢 - 系统与基础设施]() Topics: [Kubernetes](), [Encryption]() Tags: [api](), [apiserver](), [audit](), [backend](), [config](), [distributed](), [encryption](), [etcd](), [flag](), [graceful-shutdown](), [ha](), [kubernetes](), [leader-election](), [load-balancer](), [operations](), [retry](), [servers](), [upgrade](), [v1-30-3]() ### AI overview This Chinese-language operations tutorial explains kube-apiserver high availability and upgrades for Kubernetes v1.30.3. It clarifies that multiple apiserver instances run concurrently without internal leader election, sharing an etcd cluster, and covers key flags, graceful shutdown, encryption configuration, audit webhooks, upgrade checks, and version-skew guidance. ### Source excerpt kube-apiserver 高可用模式：多实例共享 etcd、无需内部 leader election；核心 flag 语义（--etcd-servers、--etcd-servers-overrides、--shutdown-delay-duration、encryption-provider-config）；与 etcd/14 的联合升级检查单；Kubernetes 版本偏差策略与 etcd 矩阵指针。 ## Using an AI Agent to Convert STL Meshes into Editable CAD DevFeed: [Using an AI Agent to Convert STL Meshes into Editable CAD]() Original publisher: [Read original article]() Author: Shy Ruparel Published: 2026-08-18T00:00:00Z Content type: article Language: en Sources: [Temporal Blog]() Topics: [AI Agent](), [stl](), [3D](), [audit trail](), [retry]() Tags: [3d](), [agent](), [ai-agent](), [audit-trail](), [generate](), [model](), [retry](), [stl](), [temporal-voices]() ### AI overview The article describes using an AI agent to convert triangulated STL meshes into editable CAD and explains why validation mattered more than an autonomous agentic loop. Temporal provided durable execution, retries, progress preservation, and an audit trail, but did not improve the agent's geometric judgment. ### Source excerpt Shy Ruparel tried an AI agent to turn STL files into editable CAD. What worked was less glamorous: generate candidates, measure them, and validate. ## I spent two months screen-sharing to ship one integration. So I rebuilt it on Temporal. DevFeed: [I spent two months screen-sharing to ship one integration. So I rebuilt it on Temporal.]() Original publisher: [Read original article]() Author: Stephen Montague Published: 2026-08-06T00:00:00Z Content type: tutorial Language: en Sources: [Temporal Blog]() Topics: [proxy](), [networking](), [legacy application](), [Messaging](), [on-prem](), [retry](), [Cloud]() Tags: [legacy-application](), [messaging](), [networking](), [on-prem](), [proxy](), [retry](), [software](), [temporal-voices]() ### AI overview The article explains how to connect a cloud application to a firewalled on-premises edge site using an egress-only proxy and a single outbound Temporal connection. It describes the operational difficulties of a legacy, one-way integration and the need for bidirectional messaging, retry handling, and testing. ### Source excerpt How to durably reach a firewalled edge site with no inbound ports: an egress-only proxy that hot-reloads config over a single outbound Temporal connection. ## Towards Designing an Execution Control System with Metastability Resilience DevFeed: [Towards Designing an Execution Control System with Metastability Resilience]() Original publisher: [Read original article]() Author: Murat (noreply@blogger.com) Published: 2026-08-02T06:46:15Z Content type: article Language: en Sources: [Metadata]() Topics: [execution](), [Resilience](), [Databases](), [scaling](), [Latency](), [retry](), [systems](), [Network]() Tags: [databases](), [execution](), [fault-tolerance](), [latency](), [metastability](), [my-paper](), [network](), [resilience](), [retries](), [retry](), [scaling](), [systems]() ### AI overview The article discusses designing a metastability-tolerant execution control system for databases. It explains how mixed workloads, overload, delayed signals, retries, and workload misclassification can create self-sustaining degraded states in cloud database systems. ### Source excerpt This week, I presented this paper at ICCCN'26. This is joint work with Aleksey Charapko (University of New Hampshire) and my MongoDB colleagues Matt Broadstone, Daniel Gomez Ferro, and Akshat Vig. The paper investigates how to build a metastability tolerant execution control system (ECS) for a database. Why? Modern databases are complex networked systems serving mixed workloads: short queries (that want an answer in milliseconds) sitting next to analytics jobs (that want the CPU for multiple seconds). The arrival rate of requests is effectively unbounded, but of course, the server's resources are not. And, unfortunately, elastic scaling does not save you here. Scaling takes minutes, whereas, overload takes seconds. Admission control tries to guard the front door (more on this later), but the component that mediates contention once requests reach the backend is the execution control system (ECS). Unlike a closed system OS scheduler, which strives for fairness and completeness by giving runtime for every thread, faced with an open environment the ECS can only afford to protect short latency-sensitive queries and shed the excess, pushing the burden of waiting back across the network to the clients. This doesn't mean that long tasks are starved, as they can retry until capacity permits execution. However, shedding load across a network is risky business. Clients do not see the server slow down, instead they time out and retry aggressively. Moreover, workloads are also unpredictable. A query that looks short may hang on a lock or blow up into a scan. The combination of delayed signals, retries, and misclassification makes the cloud databases a fertile ground for failures. The specific failure we worry about here is metastability: the system gets pushed into a degraded state, and the degraded state sustains itself even after the original trigger is removed. The mechanisms you build for resilience (the retries and the queues) turn into positive feedback loops after a trigg ## Handling Application States and Retries in Cypress End-to-End Tests DevFeed: [Handling Application States and Retries in Cypress End-to-End Tests]() Original publisher: [Read original article]() Author: Gleb Bahmutov Published: 2026-07-28T04:00:00Z Content type: tutorial Language: en Sources: [Gleb Bahmutov]() Topics: [Cypress](), [test](), [App]() Tags: [api](), [can](), [cypress](), [errors](), [flaky](), [plugin](), [products](), [retry](), [test](), [testing](), [tests]() ### AI overview This tutorial explains how to design Cypress end-to-end tests for loading, empty, and error states. It recommends detecting known error states immediately and using in-place retries with cypress-recurse when appropriate, instead of relying only on whole-test retries. ### Source excerpt Imagine a typical application that loads some data. The app starts in the initial state, starts loading, then one of 3 possible states ar ## Your SQS consumer can hang forever by default DevFeed: [Your SQS consumer can hang forever by default]() Original publisher: [Read original article]() Author: Ivan Cernja Published: 2026-07-28T00:00:00Z Content type: article Language: en Sources: [Encore Updates]() Topics: [Amazon Simple Queue Service (SQS)](), [SDK](), [Rust](), [Amazon Web Services]() Tags: [aws](), [health-check](), [queue](), [retry](), [rust](), [sdk](), [sqs](), [tcp](), [timeout]() ### AI overview The article explains that the AWS Rust SDK has no request timeout by default. A failed or silently dropped connection can leave an SQS receive waiting indefinitely, stopping a sequential consumer loop while the process and health check still appear healthy. It describes fixing the issue with a per-attempt timeout. ### Source excerpt The AWS Rust SDK ships no request timeout by default, so one SQS receive on a dead connection can hang a whole consumer with nothing in the logs. ## A Framework for Running Long-Running AI Coding Agents at Scale DevFeed: [A Framework for Running Long-Running AI Coding Agents at Scale]() Original publisher: [Read original article]() Author: Alexandre Zajac Published: 2026-06-29T15:31:27Z Content type: article Language: en Sources: [Hungry Minds]() Topics: [Artificial Intelligence](), [Loop Engineering](), [Software Engineering](), [context window](), [Orchestration](), [cursor]() Tags: [agents](), [ai](), [ai-agents](), [claude](), [coding](), [context-window](), [cursor](), [engineering](), [implementation](), [memory](), [retry]() ### AI overview A DoorDash engineer describes a framework for using AI coding agents on multi-thousand-line features. It separates research, planning, implementation, and review, while using progress files, retryable loops, and deterministic orchestration to preserve continuity across context windows. ### Source excerpt PLUS: OpenAI's custom chip 🤖, Netflix Cassandra 📊, CQRS deep dive ⚡ ## 3 Resilience Patterns That Keep a Backend Service Up When Its Dependencies Aren't DevFeed: [3 Resilience Patterns That Keep a Backend Service Up When Its Dependencies Aren't]() Original publisher: [Read original article]() Author: The T-Shaped Dev Published: 2026-06-27T10:19:01Z Content type: tutorial Language: en Sources: [The T-Shaped Dev]() Topics: [Resilience](), [Back end](), [API]() Tags: [backend](), [circuit](), [dependencies](), [patterns](), [resilience](), [retry](), [timeout]() ### AI overview A tutorial on using timeouts, retries with backoff, and circuit breakers to limit cascading failures in backend services when dependencies become slow or unreliable. It explains how unbounded waiting can exhaust server resources and describes how to layer these resilience patterns. ### Source excerpt Timeout, retry with backoff, and circuit breaker -- what each one actually does, and the order to stack them. ## Seamless Crash Protection That Symbolicates Native Crashes DevFeed: [Seamless Crash Protection That Symbolicates Native Crashes]() Original publisher: [Read original article]() Author: Shai Almog Published: 2026-06-22T00:00:00Z Content type: article Language: en Sources: [CodeName One]() Topics: [App](), [GitHub Issues](), [Cloud](), [Disk image](), [Android](), [iOS](), [Linux](), [macOS](), [Windows](), [tvOS](), [watchOS]() Tags: [android](), [cloud](), [crash](), [data](), [github-issues](), [ios](), [linux](), [macos](), [native](), [protection](), [retry](), [storage](), [watchos](), [win32](), [windows]() ### AI overview The article describes the new com.codename1.crash client, which captures crashes on devices, removes personal data before transmission, stores reports until upload succeeds, retries failed uploads, and uses cloud-side symbolication to file deduplicated reports as GitHub issues. It covers native crash support across multiple platforms and deobfuscation of Android exceptions. ### Source excerpt The new com.codename1.crash client captures crashes on device, scrubs personal data before anything is sent, retries reliably through a storage-first queue, and lets the build cloud symbolicate native crashes and file them as GitHub issues instead of emails. ## Subscription Payment Gateway: How to Pick One for Your SaaS DevFeed: [Subscription Payment Gateway: How to Pick One for Your SaaS]() Original publisher: [Read original article]() Author: Ayush Agarwal Published: 2026-06-09T00:00:00Z Content type: comparison Language: en Sources: [Dodo Payments Blog]() Topics: [Software as a service](), [stripe]() Tags: [buyers-guide](), [cancellation](), [features](), [geography](), [guide](), [merchant-of-record](), [payment-gateway](), [payments](), [retry](), [saas](), [set](), [stripe](), [subscription](), [subscriptions](), [webhooks]() ### AI overview A buyer's guide to subscription payment gateways for SaaS. It explains the minimum capabilities gateways need, including recurring charges, secure credential storage, retry logic, credential updates, and webhooks, then compares pure gateways, PSPs with native subscription billing, and merchant-of-record platforms. ### Source excerpt Subscription payment gateway buyers guide for SaaS founders. What features matter, how PSPs differ from MoRs, and how to evaluate by stage and geography. ## How to build deep research agents using Temporal and Braintrust DevFeed: [How to build deep research agents using Temporal and Braintrust]() Original publisher: [Read original article]() Author: Martin Bergman Published: 2026-06-03T00:00:00Z Content type: tutorial Language: en Sources: [Temporal Blog]() Topics: [Multi Agent Systems](), [observability](), [Pydantic](), [parallel](), [retry](), [Large Language Model]() Tags: [agents](), [build](), [how-to](), [multi-agent](), [observability](), [parallel](), [research](), [retry](), [tracing]() ### AI overview This tutorial explains a multi-agent deep research pipeline built with Temporal and Braintrust. It covers planning, query generation, parallel web search, report synthesis, and the use of Durable Execution, evals, and observability to handle timeouts, partial failures, and difficult debugging. ### Source excerpt Deep research agents are fragile in production. Here's how Temporal and Braintrust make them resilient with Durable Execution, evals, and tracing. ## Track customer loyalty points with durable workflows DevFeed: [Track customer loyalty points with durable workflows]() Original publisher: [Read original article]() Author: Cecil Phillip Published: 2026-06-03T00:00:00Z Content type: tutorial Language: en Sources: [Temporal Blog]() Topics: [account](), [business logic](), [Database](), [real-time](), [locking](), [retry](), [service](), [upgrade](), [Mobile]() Tags: [account](), [business-logic](), [customer](), [database](), [locking](), [loyalty](), [loyalty-points](), [mobile](), [process](), [query](), [real-time](), [retry](), [service](), [temporal](), [temporal-concepts](), [upgrades](), [workflows]() ### AI overview This article explains how to track customer loyalty points with durable Entity Workflows in Temporal. Each customer has a persistent workflow that maintains points, tier status, and activity history, while Signals handle accruals, Updates validate redemptions, and Queries read current balances. ### Source excerpt Track loyalty points with durable Entity Workflows in Temporal. See how to handle accruals, tier upgrades, and redemptions for each customer. ## OpenAI Style Usage Limits: Implementing Customer Quotas in Your Product DevFeed: [OpenAI Style Usage Limits: Implementing Customer Quotas in Your Product]() Original publisher: [Read original article]() Author: Ayush Agarwal Published: 2026-05-20T00:00:00Z Content type: tutorial Language: en Sources: [Dodo Payments Blog]() Topics: [Artificial Intelligence](), [Software as a service](), [OpenAI]() Tags: [ai](), [architecture](), [bug](), [how-to](), [implement](), [openai](), [protection](), [quotas](), [real-time](), [retry](), [trust](), [usage-based-billing]() ### AI overview A tutorial on implementing customer-facing usage limits for AI and SaaS products. It explains soft limits with email alerts, hard limits that stop billable work, real-time usage aggregation, clear limit messaging, and the protection these controls provide to customers and sellers. ### Source excerpt How to implement customer facing usage limits and quotas in your AI product. Soft caps, hard caps, alerts, and the OpenAI style limit experience. ## New updates to A2UI and Flutter's GenUI package DevFeed: [New updates to A2UI and Flutter's GenUI package]() Original publisher: [Read original article]() Author: Andrew Brogdon Published: 2026-05-14T14:18:22Z Content type: tutorial Language: en Sources: [Flutter - Medium]() Topics: [Flutter](), [ui](), [Protocol (disambiguation)](), [Framework](), [Large Language Model](), [Error Handling]() Tags: [android](), [architecture](), [error-handling](), [flutter](), [framework](), [generative-ai-tools](), [generative-ui](), [ios](), [language-models](), [protocol](), [retry](), [updates]() ### AI overview This migration guide explains updates to Flutter's genui package and the A2UI protocol. The package adopts A2UI v0.9, shifts from a Structured Output First to a Prompt First approach using JSON blocks in agent responses, and separates UI state and rendering, message transport, and conversation management into distinct layers. ### Source excerpt Generative UI, or GenUI for short, is a user experience pattern in which an agent not only generates content, but also makes decisions about how that content should be displayed and made interactive for the user. For Flutter developers, implementing GenUI means using A2UI, an open protocol that defines a way for agents and clients (or "renderers") to collaborate on the composition and state of a user interface. To capitalize on this, the Flutter team built genui, a package that uses A2UI to connect with an agent and provide it with a catalog of widgets to use, and then presents those widgets to the user. Both the genui package and the A2UI protocol recently got an update! The latest release of genui introduces several architectural changes to the framework. Driven by adoption of v0.9 of the A2UI protocol, this update shifts genui from a "Structured Output First" philosophy in which A2UI messages were streamed through structured output APIs to a "Prompt First" approach in which agents include blocks of JSON as text in their responses. It also decouples the architecture, providing more direct control over how your application interacts with Large Language Models (LLMs). If you are migrating an app from v0.7.0 to v0.9.0 of the genui package, this guide covers the necessary steps, from dependency cleanup to wiring up your new chat loops. Architecture decoupling In previous versions, GenUI relied on a series of classes based on ContentGenerator. These classes hid the details of prompt construction, LLM network calls, and response parsing. The latest version of package:genui removes ContentGenerator. Instead, the framework is now split into distinct layers: Engine (SurfaceController): Manages the state and rendering of your UI. Transport (A2uiTransportAdapter): Streams messages between the agent and renderer. Facade (Conversation): Provides a high-level API for managing chat states. This decoupling means you have control over chat history, retry logic, and error handling. ## Bad Retries Can Break Good Systems DevFeed: [Bad Retries Can Break Good Systems]() Original publisher: [Read original article]() Author: Raul Junco Published: 2026-05-09T11:45:53Z Content type: tutorial Language: en Sources: [System Design Classroom]() Topics: [systems](), [Back end](), [distributed-systems](), [incident](), [Latency]() Tags: [backend](), [circuit](), [gateway](), [incident](), [latency](), [retry](), [timeout]() ### AI overview A practical guide to handling transient faults in backend and distributed systems. It explains that uncontrolled retries can increase pressure on slow or overloaded dependencies and recommends backoff, jitter, retry budgets, idempotency, and circuit breakers to reduce user-facing incidents. ### Source excerpt A practical guide to using backoff, jitter, retry budgets, idempotency, and circuit breakers in backend systems. [Next page]()