---
title: How do I cut my Langfuse Cloud bill?
description: "Langfuse Cloud billing is based on ingested units. Practical ways to bring a bill down: sampling, trimming payloads, and picking the right plan."
tags: [platform, cloud]
---

# How do I cut my Langfuse Cloud bill?

Langfuse Cloud [pricing](/pricing) is based on the number of ingested units per billing period.

`Units = Traces + Observations + Scores` ([data model](/docs/administration/billable-units))

Most cost spikes result from ingesting **too many traces or overly verbose observations**.
You can **cut costs quickly by sampling fewer traces or logging only essential data**—all while preserving your core insights.

Use our [pricing calculator](/pricing?calculatorOpen=true) to estimate how different usage levels impact your monthly costs.

You can track your unit consumption in real-time via the "Langfuse Usage Management" dashboard:

  ![Langfuse Usage Management
  Dashboard](/images/docs/cloud-usage-management-dashboard.png)

## Option 1: Reduce observations per trace

Every observation within a trace counts toward your unit total. Some observations may be overly detailed or irrelevant to your specific use case. Steps to remove them:

1. Review your traces to identify low-value or unnecessary observations.
2. Update your integration/instrumentation to exclude these observations.
   - For most integrations, you define which observations are created. Thus, you can remove them by updating your instrumentation.
   - If you use the Python SDK (v3, OpenTelemetry-based), all OpenTelemetry spans are exported to Langfuse by default. If some observations are not relevant, you can [filter out observations by instrumentation scope](/docs/sdk/python/sdk-v3#filtering-by-instrumentation-scope).
   - If you use the JS SDK (v4, OpenTelemetry-based), you can [filter out irrelevant spans and observations](/docs/observability/sdk/advanced-features#filter-exported-spans).

## Option 2: Sample fewer traces

Keeping all traces is often valuable for LLM application development. Unlike traditional observability:

- Dynamic sampling based on error levels isn't feasible since you only know if a trace is interesting after completion (through user feedback, LLM-as-a-judge evaluation, etc.).
- Retaining all traces supports model distillation efforts down the line.

However, if your application operates at significant scale, sampling can be a reasonable cost-cutting strategy. Check out the [sampling docs](/docs/tracing-features/sampling) to learn more.

## Get notified before costs spike

You can [set up spend alerts](/docs/administration/spend-alerts) to receive email notifications when your Langfuse Cloud organization exceeds predefined monetary thresholds.
This helps you monitor spending and take action before unexpected costs occur.
You can configure spend alerts within your organization settings.

---

Have questions regarding your Langfuse bill? Reach out to [support](/support).

<!-- agent-instructions -->

---

## 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/faq/all/cutting-costs.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**: <https://langfuse.com/llms.txt>, 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 <resource> <action>`) to read or write traces, prompts, datasets, and scores from the terminal.

Found an error in these docs? Please open an issue at <https://github.com/langfuse/langfuse-docs/issues>.