Problem Description
The logs MCP tool consistently fails with context deadline exceeded after exactly 60 seconds, regardless of the parameters provided. Passing timeout: 120 does not extend the deadline. This makes the tool completely unusable for downloading any workflow logs.
Command/Tool
- Tool: logs
- MCP server: agenticworkflows
Steps to Reproduce
-
Run logs with any combination of parameters:
printf '{"start_date":"-1d","count":3,"max_tokens":3000}' | agenticworkflows logs .
-
Wait ~60 seconds
-
Observe error: Error [0]: context deadline exceeded
-
Retry with timeout: 120 — same 60s timeout still occurs:
printf '{"start_date":"-1d","count":3,"max_tokens":3000,"timeout":120}' | agenticworkflows logs .
-
Retry with specific workflow_name: "smoke-copilot" filter — still 60s timeout.
Expected Behavior
- Logs download successfully within reasonable time
timeout parameter should extend the deadline beyond 60s
- Filtering with
workflow_name or count: 3 should reduce response time
Actual Behavior
Every logs call exits with:
Error [0]: context deadline exceeded
after exactly 60022ms (±50ms) regardless of parameters.
Environment
- Repository: github/gh-aw
- Run ID: 30146695863
- Date: 2026-07-25
- Test variants that all failed:
{"start_date":"-1d","count":3,"max_tokens":3000} → 60022ms timeout
{"start_date":"-1d","count":3,"max_tokens":3000,"timeout":120} → 60019ms timeout
{"workflow_name":"smoke-copilot","count":3,"max_tokens":3000} → 60022ms timeout
Impact
- Severity: Critical
- Frequency: Always (100% reproduction rate across 3 distinct test variants)
- Workaround: None known — the
logs tool is completely non-functional
Additional Context
The 60s hard timeout appears to be a server-side deadline that is not respected by the timeout client parameter. The MCP keepalive pings succeed (5 pings at 10s intervals confirm the connection is alive), so this is not a network connectivity issue — the server simply does not complete the log download within the hard 60s deadline.
Generated by 🧪 Daily Cli Tools Tester · sonnet46 · 42.7 AIC · ⌖ 9.53 AIC · ⊞ 9.4K · ◷
Problem Description
The
logsMCP tool consistently fails withcontext deadline exceededafter exactly 60 seconds, regardless of the parameters provided. Passingtimeout: 120does not extend the deadline. This makes the tool completely unusable for downloading any workflow logs.Command/Tool
Steps to Reproduce
Run
logswith any combination of parameters:Wait ~60 seconds
Observe error:
Error [0]: context deadline exceededRetry with
timeout: 120— same 60s timeout still occurs:Retry with specific
workflow_name: "smoke-copilot"filter — still 60s timeout.Expected Behavior
timeoutparameter should extend the deadline beyond 60sworkflow_nameorcount: 3should reduce response timeActual Behavior
Every
logscall exits with:after exactly 60022ms (±50ms) regardless of parameters.
Environment
{"start_date":"-1d","count":3,"max_tokens":3000}→ 60022ms timeout{"start_date":"-1d","count":3,"max_tokens":3000,"timeout":120}→ 60019ms timeout{"workflow_name":"smoke-copilot","count":3,"max_tokens":3000}→ 60022ms timeoutImpact
logstool is completely non-functionalAdditional Context
The 60s hard timeout appears to be a server-side deadline that is not respected by the
timeoutclient parameter. The MCP keepalive pings succeed (5 pings at 10s intervals confirm the connection is alive), so this is not a network connectivity issue — the server simply does not complete the log download within the hard 60s deadline.