June 2026 Summaries

Temporal, a platform for building durable and scalable workflows, presents developers with a crucial decision between using timers and timeouts, which can significantly impact the robustness of their applications. Timers in Temporal are designed to implement business logic, allowing workflows to wait and then execute specific actions, while timeouts are meant to detect failures by bounding the duration of operations. Temporal's timers are durable, enabling workflows to continue from where they left off even if a worker goes down, thus supporting massive concurrency with minimal resource consumption. Conversely, timeouts are critical for managing activity failures, offering settings like Start-To-Close, Schedule-To-Close, Schedule-To-Start, and Heartbeat timeouts, each serving different purposes in failure detection and management. The article emphasizes that using workflow-level timeouts for business logic can lead to issues, as these are enforced server-side without notifying the workflow code, thereby potentially terminating executions abruptly. Instead, developers are encouraged to use timers within workflows to handle business logic, ensuring that the workflows can respond appropriately to time-limit conditions. The distinction between timers and timeouts is vital for developers to understand in order to harness Temporal's powerful time-management capabilities effectively and avoid common pitfalls.

The June newsletter highlights various updates and advancements in Temporal's offerings, including new AI capabilities and major releases such as the AWS Strands SDK integration and a Cloud CLI extension. Temporal also announced the availability of Stable IPs for Namespace Endpoints and the pre-release of Custom Roles for granular permissions in Temporal Cloud. The release of Temporal OSS v1.31 introduces features like Serverless Workers and Task Queue Priority & Fairness. The newsletter also promotes community engagement through livestreams, events, and a hands-on Python tutorial for building durable job queues. Additionally, it features insights from community contributions, such as Rapidflare's use of Temporal for processing large volumes of documents, and addresses user queries related to technical challenges, offering solutions and enhancements like upgrading SDK versions and implementing Workflow strategies.

Temporal has introduced the Temporal Cloud Billing API to enhance cost transparency and management in its usage-based billing system, providing detailed insights into cost attribution and usage optimization. The API offers Namespace- and Tag-level cost attribution with hourly granularity and integrates with Datadog Cloud Cost Management and Vantage for seamless monitoring and analysis. Additionally, the Billable Action Count metric provides minute-level usage data to aid in optimizing Workflows and forecasting expenses. These tools address long-standing customer requests for clear cost attribution and allow businesses to allocate budgets more effectively, track real-time business events, and ensure that Temporal delivers a strong return on investment. The Billing API leverages Temporal Workflows for report generation, ensuring a fair and distributed process even when data sources have strict rate limits. This approach allows customers to analyze, sort, and manage costs efficiently, fulfilling the need for a deeper understanding of where financial resources are being allocated within the platform.

Temporal Cloud introduced the metric temporal_cloud_v1_billable_action_count to track the per-second rate of billable actions by Action Type and Workflow Type, offering insights into which workflows consume the most actions and drive costs. This metric aids in cost analysis, debugging, and alerting, though it serves as an estimate rather than a precise billing tool, and not all Action Types are included. It is aggregated at the Workflow Type and Action Type level, rather than per individual Workflow Execution, allowing for trend and anomaly identification. The metric can be used to rank actions by consumption, estimate total actions over a period, and optimize system usage by identifying inefficiencies. It also facilitates debugging by providing different perspectives on Workflow execution paths and can be used to detect rare issues or regressions post-deployment. Alerts can be set up based on action volume spikes or high retry rates. Beyond immediate usage, the metric assists in usage planning, measuring optimization impacts, and attributing per-team usage, while operational reminders emphasize proper query conventions and data management.

Temporal has introduced the Billable Actions metric alongside the Billing API to provide enhanced usage visibility for their customers, transforming what began as an internal engineering initiative into a valuable tool for both internal and external users. Previously, customers could only see aggregate usage data on the Temporal Cloud dashboard, which limited their ability to identify specific Workflow Types or Action Types that were driving costs. The new metric allows for detailed breakdowns via OpenMetrics, enabling customers to query and receive time series data broken down by action type and workflow type, which aids in identifying high-volume workflows and opportunities for optimization. This granular visibility has significantly improved Temporal's internal validation processes, allowing them to more accurately verify metering and troubleshoot discrepancies with increased confidence. By releasing this feature, Temporal anticipates that customers will benefit from greater usage visibility, optimization opportunities, and improved validation and alerting capabilities, thereby enhancing their ability to manage and audit their usage independently.

The narrative recounts the experience of building a customized, conference-level registration system for Temporal's annual event, Replay, utilizing the Temporal platform. The project involved developing an intricate ecommerce system that handled payments, emails, and sensitive data while ensuring reliability and durability through Temporal's Workflows, Activities, and Signals. By leveraging Temporal's features, the development team transformed what could have been a complex integration of various PaaS/SaaS systems into a streamlined, observable process that managed everything from ticket sales to attendee registration and email notifications. This approach minimized the risk of common ecommerce issues, such as duplicate charges, by shifting responsibility to the Temporal platform, which maintained state and managed retries automatically. The author reflects on the benefits of using Temporal, particularly in terms of simplifying the architecture, enhancing error handling, and providing a coherent event history for customer support, highlighting the platform's suitability for long-lived, event-driven processes that go beyond the capabilities of traditional request/response models.

In a discussion on the implementation of agentic systems, the text explores the challenges and successes businesses face when integrating agents to transform into Intelligent Businesses. The author emphasizes the importance of using Temporal as a foundational framework for building durable, secure, and efficient agentic platforms that enhance decision-making by processing and summarizing large data volumes for human understanding and action. The text highlights the necessity of establishing proper data security practices, building AI platform teams, and ensuring user-level authentication to prevent data exposure risks. It argues that deployed agents, supported by Temporal's capabilities such as Workflow Executions and Activities, are more effective than desktop agents for specific business workflows, allowing for customized control and seamless integration into existing infrastructures. The successful transformation into an Intelligent Business involves not just the technical implementation of agents but also fostering organizational trust and adoption, with Temporal providing a robust environment for developing and evolving these systems.

Almog Baku, CEO and co-founder of Kelet AI, explains how Kelet, built on Temporal, addresses the challenge of diagnosing AI quality failures that do not manifest through traditional stack traces but rather emerge from patterns across multiple sessions. Kelet is designed to continuously diagnose these failures by using a sophisticated architecture that processes each session as it arrives, accumulates hypotheses, and reasons across the accumulated set to find root causes. The system employs Temporal's durable orchestration capabilities, such as long-running state, durable execution, event-driven coordination, and cross-workflow gating, to build a robust analysis pipeline that identifies failure clusters and generates root cause evidence. This infrastructure enables Kelet to diagnose thousands of sessions daily without human intervention, leveraging Temporal's capabilities to focus on solving AI quality issues rather than building coordination infrastructure from scratch.

Workflow Streams is a new feature for Temporal's Durable Execution that allows for the creation of rich interactive user interfaces by providing a durable streams abstraction, facilitating the seamless updating of UIs through incremental updates. This feature is beneficial for applications such as AI agents, payment processing, and order status updates by allowing monitors to intervene or confirm correct operations. The library uses Temporal's core primitives, like Workflows, Signals, Updates, and Queries, to maintain an ordered sequence of typed events that can be published and subscribed to by various components including Activities, Child Workflows, and external clients. The system ensures exactly-once semantics, enabling reliable event tracking and minimal overhead for publishing within Workflows. Available in Python and TypeScript during its Public Preview, Workflow Streams integrates with frameworks like OpenAI Agents SDK and Google ADK, allowing for responsive UI updates with adjustable batch intervals. It contrasts with traditional messaging systems like Redis by providing inherent exactly-once delivery without additional infrastructure, making it suitable for scenarios involving multiple independent streams tied to single Workflows rather than high-throughput streaming needs.

Francesco Coacci, a founding engineer at Monk, discusses the strategic migration of workflows from Inngest to Temporal in their AI-powered accounts receivable platform. The shift was necessitated by scalability issues with Inngest, which initially facilitated rapid development but eventually became cumbersome as the number of asynchronous jobs exceeded a hundred. The migration was methodically executed in stages, ensuring both systems ran concurrently through a series of reversible pull requests, each following a strict four-step process: characterizing the existing workflow, scaffolding with Temporal, transitioning operations via a feature flag, and finally removing the Inngest job post-stability. This approach allowed for continuous testing and validation, minimizing disruption and maintaining service consistency. Key challenges included managing concurrency and ensuring workflow determinism—elements that required custom solutions due to Temporal's distinct architecture. The article emphasizes the importance of incremental migration, thorough testing, and maintaining dual systems to ensure a seamless transition, ultimately resulting in improved visibility, idempotency, and operational isolation within the platform.

In response to a community suggestion that creating a Temporal Docker Desktop extension might be impossible, the author leveraged their past experience as a Developer Advocate at Docker to build a practical extension that simplifies the process of running a local Temporal server. Initially attempting to embed the Temporal UI directly within Docker Desktop, the author realized this violated design guidelines and instead created a streamlined extension that starts a Temporal server with a click, exposing necessary endpoints and providing quick access to the full Temporal UI for detailed workflow inspection. The extension, which integrates with Docker Desktop, focuses on making the setup process easy and familiar for developers, reinforcing the importance of intuitive first steps in tool adoption. The project highlights the power of community-driven innovation, transforming a casual comment into a valuable tool by combining existing resources and expertise.

This guide outlines a method for implementing a distributed locking mechanism using Temporal Workflows without relying on external databases or centralized limiters, which often introduce additional infrastructure and potential failure points. The approach is designed for managing small pools of scarce resources, like GPUs or lab equipment, where each lock or permit is represented as a short-lived child Workflow in Temporal. By utilizing Temporal's unique Workflow ID system, the method ensures atomic operations for acquiring locks and employs lease-based recovery to handle orphaned permits. This allows for horizontal scaling of lock acquisition while maintaining the integrity and synchronization of lock states with the work they gate. The guide provides a detailed implementation plan, including code examples, configuration requirements, and best practices for setting up and operating this locking system within Temporal, offering a robust solution for resource coordination in distributed systems.

A collaborative effort between Temporal, Neo4j, Auth0, and Redpanda has unveiled the complex infrastructure supporting AI agents, highlighting the essential components beneath the user interface. This initiative, centered around a Deep Research Agent developed at Temporal, illustrates the orchestration, data persistence, security, and real-time streaming necessary for AI applications. Temporal manages the workflow, ensuring durability and reliability across multiple agent interactions. Neo4j provides a knowledge graph for persistent memory and session continuity, while Auth0 secures user authentication and API interactions. Redpanda facilitates real-time event streaming for comprehensive monitoring and analytics. The integration of these technologies results in a resilient and production-ready AI system, demonstrating the importance of infrastructure in advancing AI capabilities beyond mere model performance.

Coverwatch, a San Francisco-based AI-native commercial insurance brokerage startup, utilizes Temporal to streamline and automate complex insurance workflows. The company blends AI agents with human oversight to manage risk portfolios, identify coverage gaps, and negotiate better insurance pricing. The traditional insurance brokerage process is highly operational and asynchronous, involving multiple communication channels and edge cases, making automation challenging. By adopting Temporal, Coverwatch coordinates long-running tasks and manages various handoffs between customers, brokers, underwriters, and external systems, ensuring the process is more efficient and reliable. Temporal's platform allows workflows to be modeled as durable processes, supporting immediate execution, pausing for review, and retrying when dependencies fail, thus centralizing workflow history and easing auditing and debugging. This integration has enhanced development speed and stability, eliminating the need for custom-built orchestration infrastructure and allowing the small team to focus on domain-specific logic rather than system management.

The guide explores the complexities and solutions of managing loyalty programs at scale, emphasizing the use of Temporal's Entity Workflow pattern to ensure a durable and responsive system for each customer. By providing a dedicated, persistent workflow for each customer, the system ensures real-time processing of point accruals, tier upgrades, and redemptions, eliminating the need for shared databases and cron jobs. The workflow maintains state in-memory, processes external signals, updates, and queries, and can run continuously throughout the customer's lifetime with mechanisms like Continue-As-New to manage history size. Temporal's features, including durable timers, message passing, and deterministic replay, support the architecture, offering a scalable and reliable solution that isolates failures and maintains consistency across interactions. The guide also covers operational considerations, such as event history limits and payload size constraints, and provides a detailed implementation plan that includes defining data models, implementing activities, configuring workers, and managing workflow evolution with Temporal's Worker Versioning feature.

Deep research agents, designed to generate cited, evidence-based reports from complex queries, face significant challenges due to their intricate architecture involving multiple large language model (LLM) calls, parallel web searches, and synthesis steps. These processes are prone to failures such as API timeouts, partial search successes, and nondeterministic outputs, which necessitate infrastructure-level solutions for resilience. Collaborative efforts by Braintrust and Temporal focus on enhancing these agents with Durable Execution and observability, employing multi-agent pipelines that decompose research questions, optimize search queries, execute web searches, and synthesize findings into comprehensive reports. Temporal's framework allows each agent to function as a retryable unit, ensuring continuity despite failures, and Braintrust's integration provides detailed tracing and evaluation capabilities, facilitating systematic improvements and reducing manual interventions. This approach not only addresses common failure modes but also establishes a feedback loop for continuous quality enhancement, as demonstrated in the shared Braintrust cookbook for implementation.

The text explores the development of platform control planes using Temporal, focusing on transitioning from basic infrastructure workflows to the Entity Workflow pattern, which involves one long-running workflow per managed resource. It highlights how Temporal handles complex infrastructure operations by making steps durable, observable, and recoverable, contrasting this with traditional shell scripts that lack retries, logging, and rollback capabilities. The Entity Workflow pattern is detailed as a long-running execution that maps one-to-one with a specific resource, enabling efficient management of resource lifecycles, command execution, error handling, and state tracking. The discussion emphasizes the importance of using Temporal's features like Continue-as-New to manage long-lived workflows, ensuring idempotency, and designing for durability and scalability from the outset. Additionally, it provides guidance on interacting with Entity Workflows through Temporal Client and the significance of architectural decisions in workflow design, such as when to use Activities, methods, or Child Workflows, and how to handle errors and compensations effectively. The text concludes by encouraging the integration of Infrastructure as Code (IaC) tools like Terraform and Pulumi with Temporal for enhanced orchestration capabilities.