Codex CLI cannot resume one existing thread: stream disconnected before completion with request IDs

What version of the Codex App are you using (From “About Codex” dialog)?

codex-cli 0.139.0

What subscription do you have?

Plus

What platform is your computer?

Linux 7.0.11-gentoo-x86_64-extreme x86_64 AMD Ryzen 5 5600X 6-Core Processor

What issue are you seeing?

I cannot continue one specific older/resumed Codex CLI thread. Any message sent in that thread fails before completion.

The visible errors are:

```text
Falling back from WebSockets to HTTPS transport. stream disconnected before completion: An error occurred while processing your
request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please
include the request ID [redacted] in your message.

stream disconnected before completion: An error occurred while processing your request. You can retry your request, or contact
us through our help center at help.openai.com if the error persists. Please include the request ID [redacted] in your message.

I have multiple request IDs from repeated attempts and can provide them privately to OpenAI maintainers or support if needed.

I restarted Codex and retried, but the same thread still fails. Other Codex chats appear usable, so this seems specific to this
resumed thread.

Latest request ID: 2fbf0461-60e6-4e8b-86a1-84ebc01e0aa8

I have additional request IDs from repeated attempts and can provide them privately if needed.

What steps can reproduce the bug?

1. Open Codex CLI.
2. Resume a specific older thread that had been left suspended after weekly usage limits were reached.
3. Send any message in that resumed thread, including a minimal prompt like ci sei?.
4. Codex prints the WebSocket fallback message and then fails with `stream disconnected before completion: An error occurred

while processing your request`.

The affected thread is large and important. I do not know whether the root cause is WebSocket fallback, thread size, context
compaction, or a corrupted/resume state.

Current relevant local config:

```toml
model = "gpt-5.5"
model_reasoning_effort = "medium"
cli_auth_credentials_store = "file"
suppress_unstable_features_warning = true

[features]
voice_transcription = true

What is the expected behavior?

Codex should resume the existing thread and allow me to continue the conversation.

If the thread is too large, corrupted, or cannot be processed, Codex should provide a recoverable error or a supported way to
fork, compact, repair, or export/recover the thread context.

Additional information

This may be related to existing WebSocket fallback/connectivity issues, but this case may be different because HTTPS fallback
also fails with:

```text
An error occurred while processing your request

Possibly related issues:

• https://github.com/openai/codex/issues/13041
• https://github.com/openai/codex/issues/13103
• https://github.com/openai/codex/issues/19804

I can provide a sanitized session ID or logs if maintainers tell me which files are useful and how to redact sensitive content
safely. I will not upload ~/.codex/auth.json.