Issue Station
Sign In
Tools hereToolbox

openclaw - 💡(How to fix) Fix Gateway: Event loop delay grows to 40s+ and CPU pinned at 100% on v2026.4.29 [1 comments, 2 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#76109•Fetched 2026-05-03 04:42:12
View on GitHub
Comments
1
Participants
2
Timeline
3
Reactions
2
Author
wangyan411303wangyan411303
Participants
clawsweeper[bot]clawsweeper[bot]wangyan411303wangyan411303
Timeline (top)
closed ×1commented ×1unsubscribed ×1

Gateway process on v2026.4.29 enters a dead-loop state: CPU pinned at 100%, event loop delay exceeding 40s, rendering the service unresponsive to incoming messages across all channels (WeChat, Feishu, Discord).

Root Cause

Gateway process on v2026.4.29 enters a dead-loop state: CPU pinned at 100%, event loop delay exceeding 40s, rendering the service unresponsive to incoming messages across all channels (WeChat, Feishu, Discord).

Fix Action

Fix / Workaround

[agent/embedded] startup stages totalMs=45441 stages=model-resolution:5863ms, auth:18335ms, attempt-dispatch:21155ms


- **High**: All channels (WeChat, Feishu, Discord) completely unresponsive.
- Service eventually gets OOM-killed or fails with signal by systemd.
- Downgrade to v2026.4.27 also fails (config migration incompatibility).

### Workarounds

Code Example

[diagnostic] liveness warning: reasons=event_loop_delay,event_loop_utilization,cpu 
  eventLoopDelayP99Ms=68081.9 eventLoopUtilization=1 cpuCoreRatio=1.03
  active=1 waiting=0 queued=0

[agent/embedded] prep stages totalMs=121831
  stages=core-plugin-tools:25618ms, system-prompt:38403ms, stream-setup:38550ms

[agent/embedded] startup stages totalMs=45441
  stages=model-resolution:5863ms, auth:18335ms, attempt-dispatch:21155ms
RAW_BUFFERClick to expand / collapse

Summary

Gateway process on v2026.4.29 enters a dead-loop state: CPU pinned at 100%, event loop delay exceeding 40s, rendering the service unresponsive to incoming messages across all channels (WeChat, Feishu, Discord).

Repro Steps

  1. Start gateway with openclaw gateway run (systemd-managed) on a standard Linux desktop
  2. Service starts normally, HTTP server binds, channels connect
  3. Within 1–2 minutes, CPU spikes to 100% and event loop delay grows past 40s
  4. No messages are received or processed on any channel

Reproduces consistently even with all plugins disabled (plugins.allow: []).

Expected

Gateway should remain responsive with event loop delay in normal range (< 100ms).

Actual

  • Event loop delay P99: 40,835ms → 69,927ms and climbing
  • Event loop utilization: 1.0 (fully saturated)
  • CPU core ratio: 1.025–1.048 (single-core maxed)
  • active=0, waiting=0, queued=0 — no actual message processing occurring
  • Startup trace shows extreme latency: auth resolution 18s, system prompt generation 38–39s, stream setup 38s (total prep: 115–121s)

Environment

  • OS: Linux (Ubuntu-based, x86_64)
  • Node: v24.13.0 (via nvm)
  • OpenClaw: v2026.4.29 (a448042)
  • Agent model: bailian/qwen3.5-plus
  • Channels: WeChat, Feishu (both non-functional when this occurs)

Logs/Evidence

[diagnostic] liveness warning: reasons=event_loop_delay,event_loop_utilization,cpu 
  eventLoopDelayP99Ms=68081.9 eventLoopUtilization=1 cpuCoreRatio=1.03
  active=1 waiting=0 queued=0

[agent/embedded] prep stages totalMs=121831
  stages=core-plugin-tools:25618ms, system-prompt:38403ms, stream-setup:38550ms

[agent/embedded] startup stages totalMs=45441
  stages=model-resolution:5863ms, auth:18335ms, attempt-dispatch:21155ms

Impact

  • High: All channels (WeChat, Feishu, Discord) completely unresponsive.
  • Service eventually gets OOM-killed or fails with signal by systemd.
  • Downgrade to v2026.4.27 also fails (config migration incompatibility).

Workarounds

None identified. Service must be killed and restarted, but the issue recurs on next start.

lobster-biscuit

extent analysis

TL;DR

The gateway process enters a dead-loop state due to high event loop delay and CPU utilization, likely caused by issues with the agent model or system prompt generation, and a workaround or fix is needed to prevent the service from becoming unresponsive.

Guidance

  • Investigate the bailian/qwen3.5-plus agent model and its interaction with the system prompt generation to identify potential bottlenecks or inefficiencies.
  • Review the startup trace to understand the extreme latency in auth resolution, system prompt generation, and stream setup, which may indicate a problem with the model or the environment.
  • Consider testing with a different agent model or version to isolate the issue and determine if it's specific to the current model.
  • Analyze the logs and evidence provided to identify any patterns or correlations between the event loop delay, CPU utilization, and the agent model's behavior.

Example

No code snippet is provided as the issue seems to be related to the agent model and system configuration rather than a specific code problem.

Notes

The issue may be specific to the v2026.4.29 version of OpenClaw, and downgrading to v2026.4.27 is not a viable solution due to config migration incompatibility. Further investigation is needed to determine the root cause and develop a fix or workaround.

Recommendation

Apply a workaround by testing with a different agent model or version to isolate the issue and determine if it's specific to the current model, as the current version and model combination is causing the service to become unresponsive.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#authentication issue#prompt issue#agent setup#task chaining#parallel task

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

openclaw - 💡(How to fix) Fix Gateway: Event loop delay grows to 40s+ and CPU pinned at 100% on v2026.4.29 [1 comments, 2 participants]