Changelog

Unreleased

• Update Anthropic model defaults by @joe-elliott
  • Base: claude-sonnet-4-20250514
  • Large: claude-sonnet-4-20250514

0.22.1

• feat: relax minimum Grafana version requirement by @sd2k in #724

0.22.0

• feat: add MCP StreamableHTTP server to /mcp endpoint of LLM app by @sd2k in #691
• fix: MCP isEnabled check and improve back-compatibility by @sd2k in #717

0.21.2

• fix: handle case when nested provider differs by @sd2k in #704
• feat: add log line when MCP stream can't be found by @sd2k in #703

0.21.1

• fix: revert token timeout to 10 min by @annanay25 in #698

0.21.0

• feat: export 'enabled' function from mcp module of npm package by @sd2k in #689
• feat: refresh Grafana access token before expiry in Cloud by @sd2k in #688
• refactor: add MCP struct and simplify access token usage by @sd2k in #696

0.20.1

• feat: increase OBO user auth token TTL to 30 min by @annanay25 in #684

0.20.0

• feat: remove MCP feature flag and enable MCP by default by @annanay25 in #670

0.19.3

• feat: upgrade MCP Grafana integration to v0.4.0 by @sd2k in #673

0.19.2

• feat: remove OpenAI Assistant specific code by @csmarchbanks in #663
• feat: enable LLM app to create access tokens on-behalf-of user by @annanay25 in #644

0.19.1

• feat: add Dashboard and Sift tools by @csmarchbanks in #661
• feat: allow customizing the OpenAI API path in the app config page by @sd2k in #656

0.19.0

• feat: upgrade MCP Grafana integration to v0.3.0 by @csmarchbanks in #658
• feat: add asserts tool by @xujiaxj in #657

0.18.0

• feat: add enabled flag to MCPProvider context value and fix infinite loop when MCP is not enabled by @sd2k in #648

0.17.0

• bug: Make sure at least one message has the user role for Anthropic by @edwardcqian in #629
• Update model defaults by @csmarchbanks in #632 and #635:
  • Base: gpt-4o-mini to gpt-4.1-mini
  • Large: gpt-4o to gpt-4.1

0.16.0

• chore: bump github.com/grafana/mcp-grafana to 0.2.4 by @sd2k in #618. This adds more tools to retrieve dashboards and OnCall details.
• Switch to Anthropic's OpenAI-compatible API by @gitdoluquita in #617. This PR also adds helper functions to execute tool calls with streaming endpoints.

0.15.0

• removed mentions for public preview from README.md by @Maurice-L-R in #610
• Bump a full minor version to do a release so we can publish a version out of public preview by @SandersAaronD in #612

0.14.1

• fix: improve dev sandbox prompt by @sd2k in #598
• workaround: send publish messages over Websocket by @sd2k in #601
• fix: use public API to publish, where possible by @sd2k in #603
• fix: check for >= 3 args in publish, not === 3 by @sd2k in #604
• chore: bump github.com/grafana/mcp-grafana to 0.2.2 by @sd2k in #605
• chore: add CODEOWNERS by @sd2k in #607

0.14.0

• Allow enabling of dev sandbox via jsonData by @csmarchbanks in #595
• feat: run MCP server in backend using Grafana Live by @sd2k in #574
• feat: add MCP helper functions to the @grafana/llm package by @sd2k in #590

0.13.2

• docs: Add guide for implementing new LLM providers by @gitdoluquita in #573
• Bug: duplicate healthcheck details for backwards compatiblity by @edwardcqian in #585

0.13.1

• update action versions in plugin-release by @csmarchbanks in #577
• fix: update build:all command to also build plugin frontend by @sd2k in #578

0.12.0

• Add support for OpenAI assistant functionality
• Upgrade various dependencies

0.11.0

• Update model defaults:
  • Base: gpt-3.5-turbo to gpt-4o-mini
  • Large: gpt-4-turbo to gpt-4o

0.10.9

• Update wording around OpenAI and open models
• Update LLM setup instructions

0.10.8

• Bump various dependencies to fix security issues (e.g. #446)

0.10.7

• Use non-crypto UUID generation for stream request #409

0.10.6

• Fix bug where it was impossible to fix a saved invalid OpenAI url (#405)

0.10.1

• Settings: differentiate between disabled and not configured (#350)

0.10.0

• Breaking: use base and large model names instead of small/medium/large (#334)
• Breaking: remove function calling arguments from @grafana/llm package (#343)
• Allow customisation of mapping between abstract model and provider model, and default model (#337, #338, #340)
• Make the model field optional for chat completions & chat completion stream endpoints (#341)
• Don't preload the plugin to avoid slowing down Grafana load times (#339)

0.9.1

• Fix handling of streaming requests made via resource endpoints (#326)

0.9.0

• Initial backend support for abstracted models (#315)

0.8.6

• Fix panic with stream EOF (#308)

0.8.5

• Added a displayVectorStoreOptions flag to optionally display the vector store configs

0.8.1

• Add mitigation for side channel attacks

0.7.0

• Refactors repo into monorepo together with frontend dependencies
• Creates developer sandbox for developing frontend dependencies
• Switches CI/CD to github actions

0.6.4

• Fix bug where resource calls to OpenAI would fail for Grafana managed LLMs

0.6.3

• Fix additional UI bugs
• Fix issue where health check returned true even if LLM was disabled

0.6.2

• Fix UI issues around OpenAI provider introduced in 0.6.1

0.6.1

• Store Grafana-managed OpenAI opt-in in ML cloud backend DB (Grafana Cloud only)
• Updated Grafana-managed OpenAI opt-in messaging (Grafana Cloud only)
• UI update for LLM provider selection

0.6.0

• Add Grafana-managed OpenAI as a provider option (Grafana Cloud only)

0.5.2

• Allow Qdrant API key to be configured in config UI, not just when provisioning

0.5.1

• Fix issue where temporary errors were cached, causing /health to fail permanently.

0.5.0

• Add basic auth to VectorAPI

0.4.0

• Add 'Enabled' switch for vector services to configuration UI
• Added instructions for developing with example app
• Improve health check to return more granular details
• Add support for filtered vector search
• Improve vector service health check

0.3.0

• Add Go package providing an OpenAI client to use the LLM app from backend Go code
• Add support for Azure OpenAI. The plugin must be configured to use OpenAI and provide a link between OpenAI model names and Azure deployment names
• Return streaming errors as part of the stream, with objects like {"error": "<error message>"}

0.2.1

• Improve health check endpoint to include status of various features
• Change path handling for chat completions streams to put separate requests into separate streams. Requests can pass a UUID as the suffix of the path now, but is backwards compatible with an older version of the frontend code.

0.2.0

• Expose vector search API to perform semantic search against a vector database using a configurable embeddings source

0.1.0

• Support proxying LLM requests from Grafana to OpenAI