You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Bug: /global/event SSE stream loses events on reconnect (no Last-Event-ID support) #25657
Long-running tasks (multi-agent runs emitting frequent events) silently lose UI updates when the SSE stream drops — proxy idle, browser tab backgrounded, network blip. The server keeps working, the client never catches up.
The browser's native EventSource sends Last-Event-ID automatically on reconnect, but /global/event ignores it and resubscribes fresh. Events emitted during the gap are lost. UI ends up out of sync with server state.
Plugins
oh-my-openagent
OpenCode version
dev (HEAD as of 2026-05-03; reproduces on v1.14.33 too — same code path)
Steps to reproduce
Open the web UI; start an agent task that emits frequent events (any subagent run will do).
Throttle the network in DevTools or background the tab for ~30s.
Restore. UI is missing every event from the gap window — agents continued running on the server, the client just doesn't know about it.
Screenshot and/or share link
n/a — server-side behaviour, observable in any browser hitting /global/event.
Operating System
Linux x64 (server) / macOS 15 (client)
Terminal
n/a — browser-based UI
Note re: #19584 — that issue is about an after_seq query parameter and replay-error semantics, on a different surface. This one is specifically about Last-Event-ID (the standard EventSource reconnect header) being ignored on /global/event. I have a PR ready (#25658) that adds a 1024-entry ring buffer + Last-Event-ID honoring. The same pattern can be applied to /event (instance) in a follow-up.
Thanks for updating the issue! It now meets our contributing guidelines. Your bug report is clear and well-structured with good reproduction steps and system information.
Activity
github-actions commentedon May 4, 2026
This issue doesn't fully meet our contributing guidelines.
What needs to be fixed:
Please edit this issue to address the above within 2 hours, or it will be automatically closed.
This issue might be a duplicate of existing issues. Please check:
/eventendpoint replay semantics and reconnect/recovery logic being inconsistent — same endpoint, overlapping reconnect concernsIf you believe this was flagged incorrectly, please let a maintainer know.
github-actions commentedon May 4, 2026
Thanks for updating the issue! It now meets our contributing guidelines. Your bug report is clear and well-structured with good reproduction steps and system information.
github-actions commentedon Jul 4, 2026
To stay organized issues are automatically closed after 60 days of no activity. If the issue is still relevant please open a new one.