Current state after the incident

gateway service: active/running gateway RPC: ok active tasks: 0

Observed incident window, local time redacted to minute precision

[context-overflow-diag] sessionKey=<redacted-telegram-topic-session> provider=openai-codex/gpt-5.5 source=assistantError messages=24 compactionAttempts=0 observedTokens=unknown error=Context overflow: estimated context size exceeds safe threshold during tool loop. context overflow detected (attempt 1/3); attempting auto-compaction

[session-write-lock] releasing lock held for 26601ms (max=15000ms): <openclaw-agent-sessions>/sessions.json.lock Telegram callback answer failed: 400 Bad Request: query is too old and response timeout expired or query ID is invalid exec.approval.waitDecision 37149ms exec.approval.resolve failed: unknown or expired approval id plugin.approval.resolve failed: unknown or expired approval id callback handler failed: TelegramRetryableCallbackError: GatewayClientRequestError: unknown or expired approval id

[compaction] rotated active transcript after compaction (sessionKey=<redacted-telegram-topic-session>) auto-compaction succeeded for openai-codex/gpt-5.5; retrying prompt

[diagnostic] lane wait exceeded: lane=session:<redacted-telegram-topic-session> waitedMs=115269 queueAhead=1

[context-overflow-diag] sessionKey=<redacted-telegram-topic-session> provider=openai-codex/gpt-5.5 source=assistantError messages=45 compactionAttempts=0 observedTokens=unknown error=Context overflow: estimated context size exceeds safe threshold during tool loop. [session-write-lock] releasing lock held for 21911ms (max=15000ms): <openclaw-agent-sessions>/sessions.json.lock [compaction] rotated active transcript after compaction (sessionKey=<redacted-telegram-topic-session>) auto-compaction succeeded for openai-codex/gpt-5.5; retrying prompt

[diagnostic] lane wait exceeded: lane=session:<redacted-telegram-topic-session> waitedMs=210096 queueAhead=1

User-visible agent output during the same incident stated that HEARTBEAT_OK was delivered through the ordinary answer path:

"HEARTBEAT_OK came not from normal heartbeat delivery; async command completed relay output entered ordinary agent path; model answered HEARTBEAT_OK; because not recognized as heartbeat-run it was sent as ordinary final answer."

[context-overflow-diag] sessionKey=<redacted-telegram-topic-session> provider=openai-codex/gpt-5.5 source=assistantError messages=57 compactionAttempts=0 observedTokens=unknown error=Context overflow: estimated context size exceeds safe threshold during tool loop. [session-write-lock] releasing lock held for 20517ms (max=15000ms): <openclaw-agent-sessions>/sessions.json.lock [compaction] rotated active transcript after compaction (sessionKey=<redacted-telegram-topic-session>) auto-compaction succeeded for openai-codex/gpt-5.5; retrying prompt

[agent/embedded] embedded run agent end: isError=true model=gpt-5.5 provider=openai-codex error=One of "input" or "previous_response_id"or 'prompt'or 'conversation_id' must be provided. session file repaired: rewrote 1 assistant message(s), dropped 1 blank user message(s) (<redacted-session-file>)

[diagnostic] stuck session: sessionId=<redacted-agent> sessionKey=<redacted-telegram-topic-session> state=processing age=138s queueDepth=1 [diagnostic] stuck session: sessionId=<redacted-agent> sessionKey=<redacted-telegram-topic-session> state=processing age=168s queueDepth=1 [diagnostic] stuck session: sessionId=<redacted-agent> sessionKey=<redacted-telegram-topic-session> state=processing age=198s queueDepth=1

A later user-visible agent message described the same failure class:

approved exec -> command finished -> OpenClaw tries to resume agent session -> resume/followup fails -> service text is sent to Telegram

TL;DR

The issue can be mitigated by improving the handling of internal async completion and heartbeat messages to prevent them from being delivered as normal user-visible Telegram messages during context overflow and auto-compaction.

Guidance

• Review the existing issues (#72918, #73149, #66126, #74217) to understand the broader context of the problem and potential solutions.
• Investigate the implementation of context overflow auto-compaction and approved exec/tool flow to identify why internal messages are being leaked into the user-visible chat.
• Consider adding additional logging or diagnostics to capture more information about the session state and message handling during context overflow and auto-compaction.
• Evaluate the possibility of introducing a message filtering mechanism to prevent internal heartbeat and async completion messages from being sent to the user-visible chat.

Example

No specific code snippet can be provided without more information about the implementation, but an example of a message filtering mechanism might involve checking the message type or content before sending it to the user-visible chat:

if message_type == "internal" or message_content.startswith("HEARTBEAT_OK"):
    # Handle internal message silently
else:
    # Send message to user-visible chat

Notes

The provided information suggests that the issue is related to the handling of internal messages during context overflow and auto-compaction, but more investigation is needed to determine the root cause and develop a comprehensive solution.

Recommendation

Apply a workaround to filter out internal messages from being sent to the user-visible chat, such as implementing a message filtering mechanism, until a more comprehensive solution can be developed and tested.