[
https://issues.apache.org/jira/browse/HDDS-14818?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Arafat Khan updated HDDS-14818:
-------------------------------
Description:
h2. Description
Add a Recon AI chat page where operators ask natural-language questions about
the cluster and get Markdown answers. The backend fetches Recon data and
summarizes it; the UI is a modern chat surface wired to
{{{}/api/v1/chatbot/health{}}}, {{{}/models{}}}, and {{{}/chat{}}}.
h2. What the AI does
Answers questions like cluster health, capacity, keys ({{{}/vol/bucket{}}}), OM
tasks, datanodes, and pipelines. Each question is standalone — no server-side
memory. Responses can take seconds to minutes. Answers may be wrong; users
should verify important data.
h2. What the UI does
* Gate on health: disabled / not configured / full chat
* Welcome screen with suggested prompts
* Chat bubbles, Markdown rendering (tables), copy + regenerate
* Provider + model dropdowns, send/stop, retry on error
* Session chat history until "New chat"
* Alpha label + disclaimer (independent chats, possible hallucination)
h2. Look and feel
Modern chatbot (ChatGPT-style) but clearly Recon: teal/green palette, glass
panels, rounded inputs, custom star AI mark, smooth loading states, single
scroll area, readable text/tables.
was:This JIRA adds a dedicated AI Assistant page to the Recon web dashboard a
clean, conversational chat interface where users can type questions, see the
assistant's responses, and scroll through the history of their session. There
is no need to use curl commands or API tools; the experience is fully
point-and-click. A dropdown lets users choose their preferred AI provider and
model without touching any configuration files, making the assistant practical
for both engineers and non-technical cluster administrators who simply want
answers about their cluster, fast.
> Interactive AI Assistant Web Interface
> --------------------------------------
>
> Key: HDDS-14818
> URL: https://issues.apache.org/jira/browse/HDDS-14818
> Project: Apache Ozone
> Issue Type: Sub-task
> Components: Ozone Recon
> Reporter: Arafat Khan
> Assignee: Arafat Khan
> Priority: Major
> Labels: pull-request-available
>
> h2. Description
> Add a Recon AI chat page where operators ask natural-language questions about
> the cluster and get Markdown answers. The backend fetches Recon data and
> summarizes it; the UI is a modern chat surface wired to
> {{{}/api/v1/chatbot/health{}}}, {{{}/models{}}}, and {{{}/chat{}}}.
> h2. What the AI does
> Answers questions like cluster health, capacity, keys ({{{}/vol/bucket{}}}),
> OM tasks, datanodes, and pipelines. Each question is standalone — no
> server-side memory. Responses can take seconds to minutes. Answers may be
> wrong; users should verify important data.
> h2. What the UI does
> * Gate on health: disabled / not configured / full chat
> * Welcome screen with suggested prompts
> * Chat bubbles, Markdown rendering (tables), copy + regenerate
> * Provider + model dropdowns, send/stop, retry on error
> * Session chat history until "New chat"
> * Alpha label + disclaimer (independent chats, possible hallucination)
> h2. Look and feel
> Modern chatbot (ChatGPT-style) but clearly Recon: teal/green palette, glass
> panels, rounded inputs, custom star AI mark, smooth loading states, single
> scroll area, readable text/tables.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
