Chat Thread | Zyeon UI

Preview in your theme

ready · grouping (same sender in a row shares one avatar + name header), a “Today” day separator, the unread divider the first paint lands on, file/image/link attachments, reactions — and the failed message’s Retry really re-sends (sending → sent)

Yesterday

Conversation started · Priority support

You21:32

Hey — our nightly export job started failing around 02:00.

Same stack trace on every run, so it doesn't look like a flake.

Attaching last night's log — the interesting part starts at line 812.

export-2026-07-25.log

Ada · Support21:41

Thanks, I can reproduce it. The worker is timing out while it waits on the warehouse connection pool, so the export never gets a session.

🙏1

Leaving it with the platform team overnight — I'll pick it back up first thing.

Today

You08:41

Thanks Ada. I'll check in tomorrow morning.

You09:02

Also — can we bump staging to the same pool size so we catch this before production next time?

Not delivered

Unread messages

Ada · Support09:04

Morning! Root cause confirmed: last week's release lowered the pool cap to 4, and the export needs 8 concurrent sessions.

Here's the retry window from the dashboard, plus the incident we're tracking it under.

retry-window.png incident INC-2291

👍2🎉1

The fix is deploying now — I'll confirm once tonight's 03:00 run comes back green.

streamingId · “Stream a reply” appends one assistant message and grows it chunk by chunk with a trailing caret; the view stays pinned to the newest text and never jumps. Stop halts it mid-stream; the timer is cleared on unmount

Yesterday

Conversation started · Priority support

You21:32

Hey — our nightly export job started failing around 02:00.

Same stack trace on every run, so it doesn't look like a flake.

Attaching last night's log — the interesting part starts at line 812.

export-2026-07-25.log

Ada · Support21:41

Thanks, I can reproduce it. The worker is timing out while it waits on the warehouse connection pool, so the export never gets a session.

🙏1

Leaving it with the platform team overnight — I'll pick it back up first thing.

Today

You08:41

Thanks Ada. I'll check in tomorrow morning.

You09:02

Also — can we bump staging to the same pool size so we catch this before production next time?

Not delivered

Ada · Support09:04

Morning! Root cause confirmed: last week's release lowered the pool cap to 4, and the export needs 8 concurrent sessions.

Here's the retry window from the dashboard, plus the incident we're tracking it under.

retry-window.png incident INC-2291

👍2🎉1

The fix is deploying now — I'll confirm once tonight's 03:00 run comes back green.

auto-scroll etiquette · scroll UP inside the thread first, then press “Receive a message”: you are not yanked down, a “N new messages” pill appears instead — pressing it scrolls to the bottom and clears the count. Press while already at the bottom and it just follows along

Yesterday

Conversation started · Priority support

You21:32

Hey — our nightly export job started failing around 02:00.

Same stack trace on every run, so it doesn't look like a flake.

Attaching last night's log — the interesting part starts at line 812.

export-2026-07-25.log

Ada · Support21:41

Thanks, I can reproduce it. The worker is timing out while it waits on the warehouse connection pool, so the export never gets a session.

🙏1

Leaving it with the platform team overnight — I'll pick it back up first thing.

Today

You08:41

Thanks Ada. I'll check in tomorrow morning.

You09:02

Also — can we bump staging to the same pool size so we catch this before production next time?

Not delivered

Ada · Support09:04

Morning! Root cause confirmed: last week's release lowered the pool cap to 4, and the export needs 8 concurrent sessions.

Here's the retry window from the dashboard, plus the incident we're tracking it under.

retry-window.png incident INC-2291

👍2🎉1

The fix is deploying now — I'll confirm once tonight's 03:00 run comes back green.

onLoadOlder · scroll to the TOP of the thread: 20 older messages are prepended and the message you were reading keeps its exact on-screen position. Three pages available, then the callback is dropped and the history really ends

11 messages loaded · 0 of 3 older pages fetched

Yesterday

Conversation started · Priority support

You21:32

Hey — our nightly export job started failing around 02:00.

Same stack trace on every run, so it doesn't look like a flake.

Attaching last night's log — the interesting part starts at line 812.

export-2026-07-25.log

Ada · Support21:41

Thanks, I can reproduce it. The worker is timing out while it waits on the warehouse connection pool, so the export never gets a session.

🙏1

Leaving it with the platform team overnight — I'll pick it back up first thing.

Today

You08:41

Thanks Ada. I'll check in tomorrow morning.

You09:02

Also — can we bump staging to the same pool size so we catch this before production next time?

Not delivered

Ada · Support09:04

Morning! Root cause confirmed: last week's release lowered the pool cap to 4, and the export needs 8 concurrent sessions.

Here's the retry window from the dashboard, plus the incident we're tracking it under.

retry-window.png incident INC-2291

👍2🎉1

The fix is deploying now — I'll confirm once tonight's 03:00 run comes back green.

loading · skeleton bubbles alternate sides at the same rhythm as the real thread, so nothing shifts when the history lands

empty · default empty-state slot (icon + text), overridable through emptyState

No messages yet

Say hello — the conversation will show up here.

error · onRetryLoad provided → pressing “Try again” really flips the branch to ready

Couldn't load this conversation

The message history didn't come back.

Loading preview…

This component is part of Pro

The install command, prompt and concept map are included with a Pro license — one purchase unlocks every Pro component, forever.

Unlock with Pro