COST • OpenClaw quick fix

Large transcript freezes? Move compaction off the hot path.

A fresh OpenClaw commit moved compaction planning into a bounded worker-thread path because large transcript planning can monopolize the agent event loop. For operators, the lesson is simple: when a long session freezes around history pruning or summarization, treat transcript size and compaction as a reliability/cost surface, not just a slow model. Snapshot first, reproduce with a small canary, upgrade when the fix is available, and keep expensive long-context runs behind a responsiveness check.

In simple words: Make a safe copy if the step could change your setup, try the smallest check, then confirm OpenClaw is back to normal before doing more. The source link and commands stay below for people who want the technical detail.

Checklist

Suspect compaction when long sessions freeze or become unresponsive around transcript/history pruning, especially after very large tool output or multi-agent logs.

Compaction freeze triage sheet

- Approximate transcript size / longest tool output:
- Did the freeze happen during pruning, summary, or model call?
- OpenClaw version/commit:
- Canary task that should finish quickly:
- Stop rule if the canary stalls again:

Before retrying the same expensive run, save the session/transcript reference and create a rollback point for config/model changes.

Check whether your OpenClaw build includes the compaction-planning worker fix; avoid claiming a provider/model outage until local event-loop pressure is ruled out.

Re-run a small canary with representative transcript size and a timer/heartbeat expectation before restarting a costly unattended workflow.

Cost guard before retrying a long run

openclaw --version
# Run a tiny representative task first.
# If the canary stalls around compaction, split the original job or shorten retained history before retrying.

For recurring jobs, add a stop rule: if compaction or summarization stalls once, shorten context, split the job, or lower transcript volume before spending more tokens.

Success looks like

• Large-history failures are separated from provider/model failures before money is spent on retries.

• A small canary proves the runtime is responsive before unattended or high-token work resumes.

• Long-running workflows have a transcript-size stop rule instead of repeatedly freezing at the same compaction step.

Source: OpenClaw compaction commit + X field signal · link

Before the next risky OpenClaw change

Turn this quick fix into a repeatable rollback path.

If this touches tokens, providers, upgrades, plugins, or shared-agent settings, do the free check first — then use a verified snapshot and one restore drill before trusting unattended work again.

What OpenClaw Safety Net adds

recurring-run stop rules for transcript growth and repeated stalls
rollback note before changing model/context settings to work around freezes
canary-task worksheet for checking responsiveness before scaling a job back up

Use the free fix when: you only need to unblock this one low-risk change and can write the backup label, smoke result, and rollback trigger in your own notes.

Buy when: this same risk will happen again, another person needs to repeat the fix, or you want a purchase-to-first-GREEN receipt, restore drill, and written rollback record before agents run unattended.

Do not buy yet if: you have not tried the free checklist, you need live support, or you are not ready to make a verified backup before changing a working setup.

Use OpenClaw Safety Net before the risky change →

Free fix first; buy the playbook only when the rollback path needs a repeatable first-GREEN proof record.

Have a high‑value nugget? Submit it.