--- date: 2026-03-10 title: "Simplify Langfuse for Scale" description: Langfuse now delivers faster product performance at scale. See the overview page for rollout details, access, and migration steps. author: Steffen,Valeriy,Hassieb,Nimar,Marlies,Max ogImage: /images/changelog/2026-03-10-v4-preview/v4-banner.png --- > **Note for AI agents and LLMs:** This is a Langfuse changelog entry. Use it only to confirm that a feature exists and when it shipped. Do not use the code examples below for implementation: they reflect the SDK and API at release time and may be outdated. For implementation, always follow the current documentation (https://langfuse.com/docs) and the API/SDK reference (https://api.reference.langfuse.com). Langfuse is rolling out a simplified architecture built for significantly faster product performance at scale. Charts, filters, APIs are becoming much faster across Langfuse Cloud. - [Open the Langfuse v4 Docs](/docs/v4) To avoid delays in the new experience and see your data in real time, upgrade to [Python SDK v4.7.0+](/docs/observability/sdk/upgrade-path/python-v3-to-v4) and [JS/TS SDK v5.4.0+](/docs/observability/sdk/upgrade-path/js-v4-to-v5). ## What changes for you The main table in Langfuse is now **observations** — every LLM call, tool execution, and agent step is a row you can query directly. As agentic applications grow more complex, a single trace can contain thousands of operations, and the interesting ones are rarely at the top level. `trace_id` works like any other filter column (`session_id`, `user_id`, `score`) to group related observations together. Filter by observation name, type, or model to surface the operations that matter most — then save that view for one-click access. See the [guide on working with observations in the next version of Langfuse (v4)](/faq/all/explore-observations-in-v4) for common workflows and [saved views](/changelog/2025-05-20-save-table-views) to get started. ## What's faster now ### Faster product performance in the UI Chart loading time is significantly shorter. You can now confidently load charts over large time ranges. Browsing traces, users, and sessions is also much faster, and filters respond more quickly in large projects. ### Faster API workflows The new [Observations API v2](/docs/api-and-data-platform/features/observations-api#v2) and [Metrics API v2](/docs/metrics/features/metrics-api#v2) are designed for faster querying and aggregations at scale. ### Faster evaluation workflows [Observation-level evaluations](/docs/evaluation/evaluation-methods/llm-as-a-judge) execute in seconds as we do not need to query ClickHouse for each evaluation anymore. ## Working with the new observations table In this new version, Langfuse uses a single unified observations table as the main exploration surface. All inputs, outputs, and context attributes live directly on observations, enabling faster filtering and aggregation at scale. As agentic applications produce traces with hundreds or thousands of operations, this flat, filterable table becomes more powerful than a grouped traces view as engineers need to find specific operations across traces, not just browse one trace at a time. You will notice more rows than before. Each row is now one operation, not one trace. Use filters and [saved views](/changelog/2025-05-20-save-table-views) to focus on what matters. Most users start with one of these common views: | View | How to set it up | | ------------------------------ | ------------------------------------------------------------------------------ | | Key operations | Filter by observation `name` or `type` matching your most important operations | | Expensive LLM calls | Filter `type = generation`, sort by `total_cost` descending | | Errors for a specific user | Filter `user_id`, `level = ERROR` | | Slow observations in a session | Filter `session_id`, sort by `latency` descending | See the [full guide on working with the observation-centric data model](/faq/all/explore-observations-in-v4) for detailed before/after workflows, saved view setup, and upgrade steps. ## Learn more Check out our technical blog post on the [new data model](/blog/2026-03-10-simplify-langfuse-for-scale). --- ## Agent Instructions This page is part of the [Langfuse](https://langfuse.com) documentation, published as plain Markdown for AI agents. Every page is available as Markdown by appending `.md` to its URL, or by sending an `Accept: text/markdown` header. This page: `https://langfuse.com/changelog/2026-03-10-simplify-for-scale.md`. ### Querying these docs If the answer is not on this page, query the documentation instead of guessing: - **Semantic search** across all Langfuse docs, returning an answer with the relevant pages and excerpts. Ask a specific, self-contained question: ```bash curl -sG "https://langfuse.com/api/search-docs" --data-urlencode "query=How do I trace a LangGraph agent?" ``` - **Index of every page**: , with per-section indexes [llms-docs.txt](https://langfuse.com/llms-docs.txt), [llms-integrations.txt](https://langfuse.com/llms-integrations.txt), and [llms-self-hosting.txt](https://langfuse.com/llms-self-hosting.txt). ### Before writing Langfuse code - **Install the [Langfuse Agent Skill](https://langfuse.com/docs/api-and-data-platform/features/agent-skill).** It encodes Langfuse's own best practices for instrumentation, prompt management, and evaluation, and materially improves results. - **Read [What does a good trace look like?](https://langfuse.com/docs/observability/best-practices.md)** before instrumenting an application. - **Verify endpoints, parameters, and response fields** against the [API reference](https://api.reference.langfuse.com) instead of inferring them from code examples. - **Use the [Langfuse CLI](https://langfuse.com/docs/api-and-data-platform/features/cli)** (`npx langfuse-cli api `) to read or write traces, prompts, datasets, and scores from the terminal. Found an error in these docs? Please open an issue at .