Keep what arrived when a stream dies

[elias-ba]

Description

The saved message was built entirely from Apollo's complete event. If the stream died before that arrived, everything the user had just watched appear on screen was thrown away, including workflow YAML that had already been sent.

The stream fold now keeps the text and the YAML as they arrive, and saves them if the complete event never comes. The message is marked as failed, so it reads as an interrupted answer rather than a finished one.

Job chat's suggested code can't be rescued this way. It's assembled from a tool call that only completes when the model finishes, so mid-stream it doesn't exist on either side.

Closes #__

Validation steps

1. mix test test/lightning/ai_assistant/ai_assistant_test.exs
2. Start a workflow chat against an Apollo you can kill, kill it once text has appeared, and check the message that saves has the text in it.

AI Usage

• I have used Claude Code

Pre-submission checklist

• I have performed an AI review of my code
• I have updated the changelog. (for the whole stack, in #5074)
• I have ticked a box in "AI usage" in this PR

[github-actions]

I have enough context to render the review. The PR is entirely AI-assistant reliability plumbing: partial-response preservation on stream death, a stuck-message reaper, failure-category storage/broadcast, shutdown_grace_period sizing, and detached telemetry handler IDs.

Security Review ✅

• S0 (project scoping): No new user-facing entrypoints; the reaper is a background Oban worker reading ai_chat_messages and oban_jobs internally, and the new failure_category/failure_message fields flow through the existing ai_assistant_channel.ex:39 authorize_session_access gate before broadcast (ai_assistant_channel.ex:296, message_processor.ex:316).
• S1 (authorization): N/A — no new handle_in/controller actions or policy surface; the reaper and telemetry :stop handler are non-user-invoked, and channel changes only enrich an existing authorized broadcast payload.
• S2 (audit trail): N/A — chat messages are conversational state, not project/instance configuration, so failure_category/failure_message writes (ai_assistant.ex:1113, message_processor.ex:194, stuck_message_reaper.ex:75) fall outside the S2 audit-resource list.