[deep-report] Consolidate and escalate 4 duplicate open issues: agenticworkflows logs tool ignores --timeout, always fails at 60s #48921

Description

The agenticworkflows logs MCP tool has been broken for at least 5 consecutive days (first reported 2026-07-25) with 4 separate duplicate open bug reports: #47728, #47928, #48299, #47491. Reproduced again directly during this analysis on 2026-07-29: agenticworkflows logs --start_date=-1d --count 5 --timeout 280 still fails with context deadline exceeded at exactly 60016ms elapsed, proving the timeout parameter has no effect on the actual RPC deadline. This blocks every downstream agent that depends on live log queries — the 2026-07-28 Agentic Workflow Audit (#48742) reported a full collection failure and a 22-day fleet-visibility gap directly because of this bug, and this DeepReport run hit the identical failure.

Expected Impact

Consolidating 4 duplicate issues into one canonical, well-evidenced bug report will stop future agents from re-filing duplicates and will make the growing blast radius (now blocking fleet-wide observability audits) visible enough to prioritize an actual fix, rather than another rediscovery.

Suggested Agent

A maintainer or coding agent with access to the agenticworkflows MCP server implementation (the bridge/gateway timeout handling, not the workflow layer).

Estimated Effort

Medium (1-4 hours) — likely a hardcoded 60s deadline somewhere in the bridge/session layer that ignores the caller-supplied timeout param.

Data Source

DeepReport analysis 2026-07-29; direct reproduction plus discussions #48675 (2026-07-28 DeepReport) and #48742 (Agentic Workflow Audit, 22-day visibility gap incident).

Generated by 🔬 Deep Report · age00 · 140.4 AIC · ⌖ 7.9 AIC · ⊞ 10.8K · ◷

• expires on Jul 31, 2026, 7:33 AM UTC-08:00