openclaw - 💡(How to fix) Fix [Bug]: Discord DM replies much slower after 2026.4.29 upgrade [1 comments, 2 participants]

Bug type

Regression (worked before, now fails)

Beta release blocker

No

Summary

After upgrading to OpenClaw 2026.4.29, Discord DM replies are noticeably slower, often taking about 30-45 seconds before a response appears.

Steps to reproduce

1. Run OpenClaw 2026.4.29 with Discord DM enabled on the affected Windows host.
2. Send a normal message to the configured Discord DM session.
3. Observe the time between the Discord message arriving and the assistant reply appearing.
4. Compare against the same DM workflow before the 2026.4.29 upgrade, where replies were noticeably faster.

Expected behavior

Discord DM replies should appear with roughly the same responsiveness as the previous working setup for comparable prompts, without a repeated per-turn startup/preparation delay.

Actual behavior

Discord DM replies are now noticeably slower after the 2026.4.29 upgrade. User-observed reply latency is commonly around 20-30 seconds. Startup/warmup has also been observed as much slower than expected: roughly 2-3 minutes until gateway ready, then 5-6 minutes before the assistant responds.

OpenClaw version

2026.4.29

Operating system

Windows 11 Home - Host system

Install method

No response

Model

openai-codex/gpt-5.5

Provider / routing chain

OpenClaw local Gateway -> OpenAI Codex OAuth -> openai-codex/gpt-5.5

Additional provider/model setup details

Discord direct session on the local Gateway. Runtime status during investigation showed current direct session at low context (~17%) and queue depth 0, so the visible slowdown does not look like an obviously full queue or maxed context window.

Possible lead, not confirmed: the title originally noted "possible repeated embedded-agent prep per turn" because the delay feels like some heavy per-turn preparation may be repeating rather than being cached/reused. Maintainers should treat that as a hypothesis only.

Logs, screenshots, and evidence

Evidence gathered locally before filing:

- openclaw status reports Gateway reachable locally at ws://127.0.0.1:18789, app 2026.4.29, stable channel, Gateway service running, and queue 0.
- Current session status reports model openai-codex/gpt-5.5, direct Discord DM session, context about 45k/272k (~17%), queue depth 0.
- User-observed latency note: Discord DM replies often take 20-30 seconds; startup/warmup observed around 2-3 minutes to gateway ready plus 5-6 minutes before assistant response.

Impact and severity

Affected surface: Discord DM direct conversation. Severity: Medium to High for interactive use. Frequency: Repeatedly observed after upgrade, but exact automated timing sample is not yet attached. Consequence: DM-based assistant use feels sluggish, especially for quick operational back-and-forth. Long wait times for heavier intensive operations up to 10 minutes.

Additional information

Known workaround: NOT_ENOUGH_INFO

Potential mitigation under consideration locally: reducing Discord history/context load, but this has not been confirmed as the root cause and was not applied as part of this report.