Home / Companies / DBOS / Blog / March 2026

March 2026 Summaries

4 posts from DBOS

Filter
Month: Year:
Post Summaries Back to Blog
Postgres, known for its extensive ecosystem, is more than just a popular database; it can handle data processing with triggers, schedule tasks, and integrate external data sources. A new PostgreSQL workflows client has been introduced, enabling direct workflow enqueueing from Postgres using SQL functions. This feature allows workflows to be initiated from database triggers, ensuring that each new table row can be processed individually, or as part of a transactional outbox pattern within a larger database transaction. The process utilizes a user-defined function in the system database to enqueue workflows by inserting a new row into the workflow_status table, which is then available for execution by a server. Additionally, workflows can receive messages through another function that inserts messages into a notifications table. For further exploration, resources such as documentation, a GitHub repository, and a Discord community are available to assist users in managing durable workflows with Postgres.
Mar 24, 2026 484 words in the original blog post.
Exploring the interplay of applications written in different programming languages, the blog post introduces a novel approach to bridging these languages using Postgres as a language-neutral control plane. By leveraging Postgres, workflows, messages, events, and streams can be exchanged between applications without introducing additional infrastructure like message brokers. This is achieved through a shared database schema, enabling seamless interoperability across languages, with workflows stored as entries in tables that applications in different languages can access with simple SQL statements. A critical aspect of this system is the use of a universal JSON-based serialization format for data that needs to be shared across languages, addressing the challenge of serialization compatibility. This format balances the need for interoperability with the flexibility of language-native serialization for intermediate data within workflows. By coordinating workflows through the database itself, this approach eliminates the need for additional orchestration services, thereby simplifying the integration of multi-language systems while maintaining the benefits of durability, transactions, and concurrency control that Postgres offers.
Mar 16, 2026 843 words in the original blog post.
DBOS has enhanced its workflow scheduling capabilities by introducing dynamic workflow scheduling, replacing the previous static configuration approach. The new system stores workflow schedules in a database, allowing for real-time modifications without the need to redeploy code, thus offering flexibility in managing periodic processes like data syncing and report generation. This update enables users to dynamically adjust schedules, view, pause, resume, or trigger them via a dashboard, and automatically backfill missed actions during downtime. The underlying mechanism utilizes a workflow_schedules table in Postgres, with scheduler threads polling for updates and using a unique ID as an idempotency key to ensure workflows are executed only once, maintaining distributed and fault-tolerant scheduling. To address the thundering herd problem, the system introduces jitter, spreading out database activity and reducing contention by converting potential spikes in writes into reads.
Mar 11, 2026 673 words in the original blog post.
DBOS has introduced several enhancements and integrations focused on durable AI over the past two months. Key updates include Kotlin support, the launch of configurable alerting, and the introduction of the DBOS MCP server in Conductor, alongside new integrations with OpenAI Agents SDK and Pydantic AI. DBOS Transact now offers features such as workflow patching, lineage tracking, explicit queue listening, and dynamic workflow scheduling, allowing for more flexible and reliable processing. Conductor enhancements include self-hosting capabilities, custom workflow alerting, workflow export, and a durable execution metrics dashboard for real-time monitoring. Additionally, new partnerships and templates, such as the Vercel + Supabase + DBOS template, enable streamlined and resilient background job execution. The platform aims to make durable workflows accessible and efficient, with a focus on reliability and ease of use.
Mar 03, 2026 875 words in the original blog post.