# Machine Learning **URL:** https://forum.opensearch.org/c/plugins/ml/46.md [Latest](https://forum.opensearch.org/latest.md) · [Categories](https://forum.opensearch.org/categories.md) · [Tags](https://forum.opensearch.org/tags.md) --- ## [About the Machine Learning category](https://forum.opensearch.org/t/about-the-machine-learning-category/1879)
**Author:** [@alolitas](https://forum.opensearch.org/u/alolitas)\ **Replies:** 0\ **Last updated:** [November 20, 2019, 4:32am UTC](https://forum.opensearch.org/t/about-the-machine-learning-category/1879 "2019-11-20T04:32:35Z")
This discussion category covers all things machine learning for OpenSearch. You can use this category to ask questions, discuss any issues you may be having using machine learning plugins in Open Distro for Elasticsearc… --- ## [How Can We Prevent AI Data Leakage?](https://forum.opensearch.org/t/how-can-we-prevent-ai-data-leakage/28307)
**Author:** [@rom\_c](https://forum.opensearch.org/u/rom_c)\ **Replies:** 0\ **Last updated:** [September 30, 2026, 3:57pm UTC](https://forum.opensearch.org/t/how-can-we-prevent-ai-data-leakage/28307 "2026-09-30T15:57:31Z")
AI tools are becoming part of everyday business workflows, but they can also create new data security concerns. Employees may accidentally share sensitive information such as customer details, credentials, internal docum… --- ## [Unable to deploy local custom model - No Controller is deployed](https://forum.opensearch.org/t/unable-to-deploy-local-custom-model-no-controller-is-deployed/28025)
**Author:** [@pythagoras](https://forum.opensearch.org/u/pythagoras)\ **Replies:** 9\ **Last updated:** [April 23, 2026, 11:22am UTC](https://forum.opensearch.org/t/unable-to-deploy-local-custom-model-no-controller-is-deployed/28025 "2026-04-23T11:22:57Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): Version 3.6 Describe the issue: I cannot not deploy a local pretrained or local custom model. I am running an OpenSeach Cluster with 3 Nodes inside K8s Clus… --- ## [Registered agent using gemini returns API key not valid](https://forum.opensearch.org/t/registered-agent-using-gemini-returns-api-key-not-valid/27957)
**Author:** [@gstorme](https://forum.opensearch.org/u/gstorme)\ **Replies:** 2\ **Last updated:** [March 30, 2026, 9:02pm UTC](https://forum.opensearch.org/t/registered-agent-using-gemini-returns-api-key-not-valid/27957 "2026-03-30T21:02:43Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): 3.5.0 Describe the issue: Following the documentation on Register agent - OpenSearch Documentation I have registered an agent: POST /\_plugins/\_ml/agents/\_r… --- ## [How to create semantic search agent using flow type](https://forum.opensearch.org/t/how-to-create-semantic-search-agent-using-flow-type/27888)
**Author:** [@YukiS](https://forum.opensearch.org/u/YukiS)\ **Replies:** 1\ **Last updated:** [March 23, 2026, 11:14pm UTC](https://forum.opensearch.org/t/how-to-create-semantic-search-agent-using-flow-type/27888 "2026-03-23T23:14:54Z")
The agent created below is a conversational type agent. If there is a way to create a flow type agent capable of both keyword search and semantic search, please let me know. --- ## [Need OpenSearch MCP example](https://forum.opensearch.org/t/need-opensearch-mcp-example/27424)
**Author:** [@asfoorial](https://forum.opensearch.org/u/asfoorial)\ **Replies:** 3\ **Last updated:** [December 16, 2025, 5:59am UTC](https://forum.opensearch.org/t/need-opensearch-mcp-example/27424 "2025-12-16T05:59:16Z")
Hi all, I am using OS 3.3.1 and I am trying to create an MCP tool that runs the following query against a specific index. Query: GET my\_pdf\_docs/\_search { “query”: { “query\_string”: { “default\_field”: “\*”, “query… --- ## [Error communicating with remote model - SSL handshake exception](https://forum.opensearch.org/t/error-communicating-with-remote-model-ssl-handshake-exception/27368)
**Author:** [@pablo](https://forum.opensearch.org/u/pablo)\ **Replies:** 10\ **Last updated:** [December 15, 2025, 4:17am UTC](https://forum.opensearch.org/t/error-communicating-with-remote-model-ssl-handshake-exception/27368 "2025-12-15T04:17:40Z")
This issue was created in the OpenSearch Security Slack channel. Hi, I am trying to connect to a privately hosted llm from OpenSearch by following the documentation Connecting to externally hosted models - OpenSearch … --- ## [Semantic field does not work with remote ada-embedding model](https://forum.opensearch.org/t/semantic-field-does-not-work-with-remote-ada-embedding-model/27614)
**Author:** [@sousu](https://forum.opensearch.org/u/sousu)\ **Replies:** 0\ **Last updated:** [December 12, 2025, 7:02pm UTC](https://forum.opensearch.org/t/semantic-field-does-not-work-with-remote-ada-embedding-model/27614 "2025-12-12T19:02:50Z")
Versions: 3.3 Describe the issue: I am getting an error when creating a semantic filed with a remote model based on this guide. Configuration: { "name": "OpenAI embedding model", "model\_group\_id": "XXXX", "algor… --- ## [Question on Memory Plugin Access Control and DLS/FLS Integration](https://forum.opensearch.org/t/question-on-memory-plugin-access-control-and-dls-fls-integration/27601)
**Author:** [@bwa](https://forum.opensearch.org/u/bwa)\ **Replies:** 1\ **Last updated:** [December 10, 2025, 6:42am UTC](https://forum.opensearch.org/t/question-on-memory-plugin-access-control-and-dls-fls-integration/27601 "2025-12-10T06:42:36Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): Describe the issue: Configuration: Relevant Logs or Screenshots: While reviewing the OpenSearch Memory plugin code, I noticed that access control seems li… --- ## [Why does my Azure OpenAI deployment occasionally return “429 Too Many Requests” even when I am under the documented rate limits?](https://forum.opensearch.org/t/why-does-my-azure-openai-deployment-occasionally-return-429-too-many-requests-even-when-i-am-under-the-documented-rate-limits/27579)
**Author:** [@Allen\_09876](https://forum.opensearch.org/u/Allen_09876)\ **Replies:** 0\ **Last updated:** [December 3, 2025, 11:07am UTC](https://forum.opensearch.org/t/why-does-my-azure-openai-deployment-occasionally-return-429-too-many-requests-even-when-i-am-under-the-documented-rate-limits/27579 "2025-12-03T11:07:39Z")
I am using the Python SDK to access Azure OpenAI (GPT-4o / GPT-4o-mini). My usage logs show that I’m well below the Tokens-per-Minute and Requests-per-Minute limits for my instance. Even so, I sometimes get: 429 Too … --- ## [Error on connecting external connector and embedding model](https://forum.opensearch.org/t/error-on-connecting-external-connector-and-embedding-model/27487)
**Author:** [@joaoolavo](https://forum.opensearch.org/u/joaoolavo)\ **Replies:** 6\ **Last updated:** [November 18, 2025, 6:45pm UTC](https://forum.opensearch.org/t/error-on-connecting-external-connector-and-embedding-model/27487 "2025-11-18T18:45:06Z")
Versions: OpenSearch 2.19 (AWS managed servers) Describe the issue: I’ve configured a connector (that goes on company’s internal API embedding model) and a model to use it. Running a predict on model works just fine (g… --- ## [Use cases and general feedback for Agentic Search](https://forum.opensearch.org/t/use-cases-and-general-feedback-for-agentic-search/27488)
**Author:** [@owaiskazi19](https://forum.opensearch.org/u/owaiskazi19)\ **Replies:** 0\ **Last updated:** [November 14, 2025, 7:36pm UTC](https://forum.opensearch.org/t/use-cases-and-general-feedback-for-agentic-search/27488 "2025-11-14T19:36:07Z")
With the recent release of OpenSearch 3.3, we introduced a new capability called Agentic Search. Posting here to gather your use cases, experiences, and any feedback you may have on this feature. Looking forward to your… --- ## [Open Telemetry trace generation from Agentic Search](https://forum.opensearch.org/t/open-telemetry-trace-generation-from-agentic-search/27411)
**Author:** [@indranilr](https://forum.opensearch.org/u/indranilr)\ **Replies:** 1\ **Last updated:** [November 14, 2025, 7:20pm UTC](https://forum.opensearch.org/t/open-telemetry-trace-generation-from-agentic-search/27411 "2025-11-14T19:20:55Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): OpenSearch 3.2 and above Describe the issue: Agentic Search needs to support OpenTelemtry based trace generation so that agent interactions can be monitored … --- ## [Cohere connection failing](https://forum.opensearch.org/t/cohere-connection-failing/27435)
**Author:** [@tejashu](https://forum.opensearch.org/u/tejashu)\ **Replies:** 4\ **Last updated:** [November 13, 2025, 6:40pm UTC](https://forum.opensearch.org/t/cohere-connection-failing/27435 "2025-11-13T18:40:49Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): Opensearch -3.2.0 Describe the issue: I am trying to connect to cohere using the connector : { “name”: “Cohere Chat Model”, “description”: “The connecto… --- ## [Anomaly detector custom expressions with time range](https://forum.opensearch.org/t/anomaly-detector-custom-expressions-with-time-range/27373)
**Author:** [@apaws06](https://forum.opensearch.org/u/apaws06)\ **Replies:** 6\ **Last updated:** [November 5, 2025, 11:34pm UTC](https://forum.opensearch.org/t/anomaly-detector-custom-expressions-with-time-range/27373 "2025-11-05T23:34:45Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): v2.19 Describe the issue: I would like to write a Custom expression in the “Anomaly Detection” to detect users login from outside working hours. The logic… --- ## [MCP Example with Streamable HTTP Not Working](https://forum.opensearch.org/t/mcp-example-with-streamable-http-not-working/27430)
**Author:** [@asfoorial](https://forum.opensearch.org/u/asfoorial)\ **Replies:** 1\ **Last updated:** [November 5, 2025, 2:20pm UTC](https://forum.opensearch.org/t/mcp-example-with-streamable-http-not-working/27430 "2025-11-05T14:20:32Z")
Hi all, I am trying the latest MCP built-in server in OS 3.3.1 and particularily the below code. It fails giving an error “httpx.ResponseNotRead: Attempted to access streaming response content, without having called \`rea… --- ## [Derived Source error with semantic field type](https://forum.opensearch.org/t/derived-source-error-with-semantic-field-type/27404)
**Author:** [@asfoorial](https://forum.opensearch.org/u/asfoorial)\ **Replies:** 1\ **Last updated:** [November 3, 2025, 1:21pm UTC](https://forum.opensearch.org/t/derived-source-error-with-semantic-field-type/27404 "2025-11-03T13:21:12Z")
OS 3.3.1 Hi all, I am getting the below error when created an index with derived\_source enabled and having a semantic field type. Your help is highly appreciated. HTTP Request: PUT test\_semantic2 { “settings”: { “in… --- ## [Embedding with nested Vs flat](https://forum.opensearch.org/t/embedding-with-nested-vs-flat/27403)
**Author:** [@tejashu](https://forum.opensearch.org/u/tejashu)\ **Replies:** 0\ **Last updated:** [October 31, 2025, 8:57am UTC](https://forum.opensearch.org/t/embedding-with-nested-vs-flat/27403 "2025-10-31T08:57:29Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): Opensearch -3.3 Describe the issue: The embedding mapping with nested like: “VECTOR\_Nested”: { “type”: “nested”, “properties”: { “embedding”: { “type”… --- ## [Best AI/NLP model for converting natural language to OpenSearch queries](https://forum.opensearch.org/t/best-ai-nlp-model-for-converting-natural-language-to-opensearch-queries/27401)
**Author:** [@jaspher](https://forum.opensearch.org/u/jaspher)\ **Replies:** 0\ **Last updated:** [October 31, 2025, 5:01am UTC](https://forum.opensearch.org/t/best-ai-nlp-model-for-converting-natural-language-to-opensearch-queries/27401 "2025-10-31T05:01:17Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):2.17.1 Describe the issue: Hi all, I’m exploring ways to generate valid OpenSearch queries (JSON format) from natural language text. \*\*Goal:\*\* Users type … --- ## [Need example that use VectorDBTool and MCP](https://forum.opensearch.org/t/need-example-that-use-vectordbtool-and-mcp/27390)
**Author:** [@asfoorial](https://forum.opensearch.org/u/asfoorial)\ **Replies:** 1\ **Last updated:** [October 30, 2025, 3:32pm UTC](https://forum.opensearch.org/t/need-example-that-use-vectordbtool-and-mcp/27390 "2025-10-30T15:32:09Z")
Hi all, I highly appreciate help to register an MCP VectorDBTool tool. Thanks --- ## [How to deploy ibm-granite/granite-4.0-h-350m model in openseach](https://forum.opensearch.org/t/how-to-deploy-ibm-granite-granite-4-0-h-350m-model-in-openseach/27389)
**Author:** [@asfoorial](https://forum.opensearch.org/u/asfoorial)\ **Replies:** 0\ **Last updated:** [October 30, 2025, 10:08am UTC](https://forum.opensearch.org/t/how-to-deploy-ibm-granite-granite-4-0-h-350m-model-in-openseach/27389 "2025-10-30T10:08:47Z")
Hi all, I was wondering how to deploy ibm-granite/granite-4.0-h-350m (granite nano family) on OpenSearch ML nodes. Granite nano has shown remarkable results in both speed and accuracy at tasks such function calling and … --- ## [Neural Sparse Indexing VS Dense Vector Indexing](https://forum.opensearch.org/t/neural-sparse-indexing-vs-dense-vector-indexing/27384)
**Author:** [@tejashu](https://forum.opensearch.org/u/tejashu)\ **Replies:** 0\ **Last updated:** [October 29, 2025, 5:18am UTC](https://forum.opensearch.org/t/neural-sparse-indexing-vs-dense-vector-indexing/27384 "2025-10-29T05:18:19Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): OpenSearch-3.2.0 Describe the issue: I ran an index with 1.2M documents . The neural sparse encoding took around 55 minutes to index the data with 240MB s… --- ## [Exclude source in Dense vector embedding increases the storage size](https://forum.opensearch.org/t/exclude-source-in-dense-vector-embedding-increases-the-storage-size/27383)
**Author:** [@tejashu](https://forum.opensearch.org/u/tejashu)\ **Replies:** 2\ **Last updated:** [October 29, 2025, 5:10am UTC](https://forum.opensearch.org/t/exclude-source-in-dense-vector-embedding-increases-the-storage-size/27383 "2025-10-29T05:10:50Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): Opensearch- 3.2.0 Describe the issue: Also the \_source excludes if set in mappings is increasing the storage size by 2X. Is this behavior expected or is t… --- ## [Feature Request: Agent DQLTool](https://forum.opensearch.org/t/feature-request-agent-dqltool/27285)
**Author:** [@wdevlin](https://forum.opensearch.org/u/wdevlin)\ **Replies:** 2\ **Last updated:** [October 13, 2025, 6:42pm UTC](https://forum.opensearch.org/t/feature-request-agent-dqltool/27285 "2025-10-13T18:42:11Z")
Versions 3.2+ OS: Describe the Feature: Feature request for Agent Tool that is an expert with DQL. Since OS Dashboards support DQL the Agent connected to Chat Assistant gains the capability of formulating precise DQL qu… --- ## [Problem with plugins ml model](https://forum.opensearch.org/t/problem-with-plugins-ml-model/27234)
**Author:** [@vio](https://forum.opensearch.org/u/vio)\ **Replies:** 2\ **Last updated:** [October 8, 2025, 11:49am UTC](https://forum.opensearch.org/t/problem-with-plugins-ml-model/27234 "2025-10-08T11:49:10Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): Opensearch version 3.2.0/ Server from Mittwald/sending requests using Postman version: 11.64.7 Describe the issue: Hello, first off I’m not a programmer so b… --- ## [Offline deployment of pretrained ML models in OpenSearch 2.18.0 — resolving DJL JNI library loading failure](https://forum.opensearch.org/t/offline-deployment-of-pretrained-ml-models-in-opensearch-2-18-0-resolving-djl-jni-library-loading-failure/27217)
**Author:** [@ark202](https://forum.opensearch.org/u/ark202)\ **Replies:** 2\ **Last updated:** [October 4, 2025, 1:16am UTC](https://forum.opensearch.org/t/offline-deployment-of-pretrained-ml-models-in-opensearch-2-18-0-resolving-djl-jni-library-loading-failure/27217 "2025-10-04T01:16:18Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): OpenSearch: 2.18.0 ML Commons Plugin: 2.18.0.0 Server OS: Linux (x86\_64) Python: 3.9 PyTorch: 2.2.2 DJL: 0.28.0 Describe the issue: While deploying pretra… --- ## [Deploying custom TEXT\_EMBEDDING model fails with "model content changed" despite matching SHA-256](https://forum.opensearch.org/t/deploying-custom-text-embedding-model-fails-with-model-content-changed-despite-matching-sha-256/27156)
**Author:** [@PetX](https://forum.opensearch.org/u/PetX)\ **Replies:** 0\ **Last updated:** [September 26, 2025, 11:14am UTC](https://forum.opensearch.org/t/deploying-custom-text-embedding-model-fails-with-model-content-changed-despite-matching-sha-256/27156 "2025-09-26T11:14:57Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): OpenSearch Dashboards: v3.1.0 Brownser : Chrome Describe the issue: When trying to deploy a custom model on OpenSearch Dashboard DevTools, using \_deploy … --- ## [Anomaly Detection Data Retention](https://forum.opensearch.org/t/anomaly-detection-data-retention/27063)
**Author:** [@jddelg](https://forum.opensearch.org/u/jddelg)\ **Replies:** 1\ **Last updated:** [September 25, 2025, 1:00pm UTC](https://forum.opensearch.org/t/anomaly-detection-data-retention/27063 "2025-09-25T13:00:50Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): 2.17.1 Hi all, Looking for clarification on the anomaly detection implementation. My logs are retained for 30 days, so I am unsure if the model is still ‘tr… --- ## [LiteLLM Connector Support](https://forum.opensearch.org/t/litellm-connector-support/27123)
**Author:** [@indranilr](https://forum.opensearch.org/u/indranilr)\ **Replies:** 1\ **Last updated:** [September 24, 2025, 10:15pm UTC](https://forum.opensearch.org/t/litellm-connector-support/27123 "2025-09-24T22:15:20Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): 3.2 Describe the issue: Add LiteLLM Connector Support as an out of box ML connector Connectors - OpenSearch Documentation . For enterprise deployments LiteLL… --- ## [Unable to configure proxy with authentication for ML model download](https://forum.opensearch.org/t/unable-to-configure-proxy-with-authentication-for-ml-model-download/26975)
**Author:** [@MHR](https://forum.opensearch.org/u/MHR)\ **Replies:** 5\ **Last updated:** [September 24, 2025, 3:08pm UTC](https://forum.opensearch.org/t/unable-to-configure-proxy-with-authentication-for-ml-model-download/26975 "2025-09-24T15:08:51Z")
Versions (relevant - OpenSearch/Dashboard/Server OS/Browser): Describe the issue: Hi, I’m trying to register ML models in OpenSearch, but my environment requires going through an HTTP proxy with authentication (user/p… [Next page](https://forum.opensearch.org/c/plugins/ml/46.md?page=1)