Bug type
Behavior bug (incorrect output/state without crash)
Beta release blocker
No
Summary
Embedded runner startup tracing can report multi-second attempt-dispatch spans without enough subphase detail to identify whether the delay came from workspace setup, prompt assembly, runtime-plan setup, or final dispatch.
Steps to reproduce
NOT_ENOUGH_INFO
Expected behavior
When the embedded runner emits a slow startup summary for phase=attempt-dispatch, the summary should include subspans that identify the first-attempt dispatch setup phases.
Actual behavior
A local trace snapshot contained 7 lines matching attempt-dispatch, including a representative line showing attempt-dispatch:11632ms, but the available evidence did not include enough dispatch subphase detail to isolate the slow setup step.
OpenClaw version
NOT_ENOUGH_INFO
Operating system
NOT_ENOUGH_INFO
Install method
NOT_ENOUGH_INFO
Model
NOT_ENOUGH_INFO
Provider / routing chain
NOT_ENOUGH_INFO
Additional provider/model setup details
NOT_ENOUGH_INFO
Logs, screenshots, and evidence
Trace/proof:
- Representative previous line in this snapshot:
"[trace:embedded-run] ... attempt-dispatch:11632ms"
Observed count: 7 lines matching "attempt-dispatch".
Likely source: embedded runner attempt dispatch / trace timing.
Impact and severity
Affected users/systems/channels: NOT_ENOUGH_INFO
Severity: Low/Medium observability gap; the observed trace shape makes slow embedded-run dispatch hard to diagnose.
Frequency: 7 observed matching trace lines in the local evidence snapshot.
Consequence: Dispatch latency can add several seconds before the model/tool attempt starts, and the current trace output does not identify the responsible setup subphase.
Additional information
Current code on origin/main emits a startup summary at src/agents/pi-embedded-runner/run.ts for phase=attempt-dispatch, while the lower-level attempt path already has separate prep/core-plugin-tool stage summaries. The bug is the missing breakdown inside the outer first-attempt dispatch region before runEmbeddedAttemptWithBackend starts.
Bug type
Behavior bug (incorrect output/state without crash)
Beta release blocker
No
Summary
Embedded runner startup tracing can report multi-second
attempt-dispatchspans without enough subphase detail to identify whether the delay came from workspace setup, prompt assembly, runtime-plan setup, or final dispatch.Steps to reproduce
NOT_ENOUGH_INFO
Expected behavior
When the embedded runner emits a slow startup summary for
phase=attempt-dispatch, the summary should include subspans that identify the first-attempt dispatch setup phases.Actual behavior
A local trace snapshot contained 7 lines matching
attempt-dispatch, including a representative line showingattempt-dispatch:11632ms, but the available evidence did not include enough dispatch subphase detail to isolate the slow setup step.OpenClaw version
NOT_ENOUGH_INFO
Operating system
NOT_ENOUGH_INFO
Install method
NOT_ENOUGH_INFO
Model
NOT_ENOUGH_INFO
Provider / routing chain
NOT_ENOUGH_INFO
Additional provider/model setup details
NOT_ENOUGH_INFO
Logs, screenshots, and evidence
Impact and severity
Affected users/systems/channels: NOT_ENOUGH_INFO
Severity: Low/Medium observability gap; the observed trace shape makes slow embedded-run dispatch hard to diagnose.
Frequency: 7 observed matching trace lines in the local evidence snapshot.
Consequence: Dispatch latency can add several seconds before the model/tool attempt starts, and the current trace output does not identify the responsible setup subphase.
Additional information
Current code on
origin/mainemits a startup summary atsrc/agents/pi-embedded-runner/run.tsforphase=attempt-dispatch, while the lower-level attempt path already has separate prep/core-plugin-tool stage summaries. The bug is the missing breakdown inside the outer first-attempt dispatch region beforerunEmbeddedAttemptWithBackendstarts.