Configuring Agentic Investigation and Chatbot Features in OpenSearch: A Complete Guide #4056
Conversation
Signed-off-by: Jiaru Jiang <[email protected]>
|
Thank you for submitting a blog post! The blog post review process is: Submit a PR -> (Optional) Peer review -> Doc review -> Editorial review -> Marketing review -> Published. |
|
Hi @PauiC, It looks like you're adding a new blog post but don't have an issue mentioned. Please link this PR to an open issue using one of these keywords in the PR description:
If an issue hasn't been created yet, please create one and then link it to this PR. |
|
@PauiC - Please open a blog issue for this so it can be published. |
|
@kolchfa-aws - Adding you for review |
|
@PauiC Let me know when your team has reviewed the draft and it's ready for my review. Thanks! |
_posts/2026-01-09-configuring-investigation-and-chatbot-features-in-opensearch.md
Outdated
Show resolved
Hide resolved
_posts/2026-01-09-configuring-investigation-and-chatbot-features-in-opensearch.md
Outdated
Show resolved
Hide resolved
_posts/2026-01-09-configuring-investigation-and-chatbot-features-in-opensearch.md
Show resolved
Hide resolved
_posts/2026-01-09-configuring-investigation-and-chatbot-features-in-opensearch.md
Show resolved
Hide resolved
PauiC
changed the title
Signed-off-by: Jiaru Jiang <[email protected]>
| ```bash | ||
| git clone https://github.com/opensearch-project/OpenSearch-Dashboards.git | ||
| cd OpenSearch-Dashboards/plugins/ | ||
| git clone https://github.com/opensearch-project/dashboards-investigation.git | ||
| git clone https://github.com/opensearch-project/dashboards-assistant.git | ||
| cd .. | ||
| yarn osd bootstrap | ||
| yarn start --no-base-path | ||
| ``` |
There was a problem hiding this comment.
For this part, with 3.5 been released, we could use 3.5 distribution directly instead of building from source
There was a problem hiding this comment.
Where can I find the latest release? I noticed that the latest version of ml on GitHub is 3.3.0.0
| # Install all built plugins into OpenSearch | ||
| cd build/distribution/local/opensearch-3.4.0-SNAPSHOT | ||
| ./bin/opensearch-plugin install file:///path/to/job-scheduler.zip | ||
| ./bin/opensearch-plugin install file:///path/to/ml-commons.zip | ||
| ./bin/opensearch-plugin install file:///path/to/skills.zip | ||
| ./bin/opensearch-plugin install file:///path/to/sql.zip | ||
| ./bin/opensearch-plugin install file:///path/to/transport-reactor-netty4.zip | ||
| ./bin/opensearch-plugin install file:///path/to/arrow-flight-rpc.zip | ||
| ./bin/opensearch # Start OpenSearch with all plugins | ||
| ``` |
There was a problem hiding this comment.
same here, we can use 3.5 distribution. but for transport-reactor-netty4 and arrow-flight-rpc, we need to build from source
| plugins.ml_commons.stream_enabled: true | ||
| plugins.ml_commons.ag_ui_enabled: true | ||
| plugins.ml_commons.mcp_connector_enabled: true | ||
| plugins.ml_commons.trusted_connector_endpoints_regex: ["^https?://localhost.*", "^https?://127\.0\.0\.1.*"] |
There was a problem hiding this comment.
we should include LLM endpoints like claude/openai
| opensearch.experimental.feature.transport.stream.enabled: true | ||
|
|
||
| # HTTP transport type | ||
| http.type: reactor-netty4 |
There was a problem hiding this comment.
for security enabled case, this should be reactor-netty4-secure
There was a problem hiding this comment.
I can only use reactor-netty4 for local building, and an error will occur if I use the reactor-netty4-secureconfiguration.
| git clone https://github.com/opensearch-project/opensearch-mcp-server-py.git | ||
| cd opensearch-mcp-server-py | ||
| pip install -r requirements.txt # Install Python dependencies | ||
| # Configure and start MCP server with your OpenSearch credentials |
There was a problem hiding this comment.
same here, guess we don't need to build from source
| "action_type": "predict", | ||
| "method": "POST", | ||
| "url": "https://bedrock-runtime.${parameters.region}.amazonaws.com/model/${parameters.model}/converse-stream", | ||
| "request_body": "{\"messages\": [${parameters._chat_history:-}{\"role\":\"user\",\"content\":[{\"text\":\"${parameters.prompt}\"}]}${parameters._interactions:-}]${parameters.tool_configs:-}}" |
There was a problem hiding this comment.
{\"system\": [{\"text\": \"${parameters.system_prompt}\"}],\"messages\": [${parameters._chat_history:-}{\"role\":\"user\",\"content\":[{\"text\":\"${parameters.prompt}\"}]}${parameters._interactions:-}]${parameters.tool_configs:-}}
There was a problem hiding this comment.
can we have latest screenshot
There was a problem hiding this comment.
maybe a screenshot for trigger investigation from chatbot as well?
| "mcp_connectors": "[{\"mcp_connector_id\":\"{% raw %}{{mcp_connector_id}}{% endraw %}\"}]" | ||
| }, | ||
| "tools": [ | ||
| { | ||
| "type": "ListIndexTool" | ||
| }, | ||
| { | ||
| "type": "SearchIndexTool" | ||
| }, | ||
| { | ||
| "type": "IndexMappingTool" | ||
| } | ||
| ], | ||
| "app_type": "os_chat" | ||
| } |
There was a problem hiding this comment.
those tools are duplicate with MCP tools
There was a problem hiding this comment.
Are there any tools that MCP doesn't have? Or should I just modify it to:
"tools": [
],
| "description": "Connector to MCP server", | ||
| "version": "1", | ||
| "protocol": "mcp_streamable_http", | ||
| "url": "http://localhost:8080" |
There was a problem hiding this comment.
missing parameters for connector?
"endpoint": "/mcp/"
}```
| { | ||
| "name": "Claude 4.5", | ||
| "function_name": "remote", | ||
| "description": "openai model", |
There was a problem hiding this comment.
nit: let's update this description
| } | ||
| ] | ||
| }, | ||
| "tools": [ |
There was a problem hiding this comment.
Don't think we need these built-in tools as they are provided by MCP server
| OPENSEARCH_URL="<your_opensearch_endpoint>" \ | ||
| OPENSEARCH_USERNAME="<your_username>" \ | ||
| OPENSEARCH_PASSWORD="<your_password>" \ | ||
| python -m src.mcp_server_opensearch --transport stream --host 0.0.0.0 --port 8080 |
There was a problem hiding this comment.
+1 to the comment above, we can replace this line with uvx opensearch-mcp-server-py --transport stream --port 8080 and skip the clone and install steps
6 participants
Description
Learn how to configure OpenSearch chatbot and investigation features, including OpenSearch backend, Dashboards frontend, and MCP server configuration.
Issues Resolved
[List any issues this PR will resolve]
Check List
By submitting this pull request, I confirm that my contribution is made under the terms of the BSD-3-Clause License.