fix(membership): ignore health monitor stop cancellation

[ReubenBond]

Problem

SiloHealthMonitor can be stopped while a timed-out direct probe is awaiting the local stall detector. The monitor-owned cancellation then reaches the generic run-loop handler and is logged as an error, causing rolling-upgrade tests which assert clean runtime logs to fail.

Solution

Classify OperationCanceledException after the monitor's stopping token is canceled as expected lifecycle completion. Preserve error logging for all other failures, including failures which occur after stop is requested.

Add focused coverage for cancellation during stall detection, ordinary unexpected run-loop failures, and unexpected stall-detector failures after stop.

Fixes #11408

Microsoft Reviewers: Open in CodeFlow

[Copilot]

Copilot review overview

🟢 Approval recommended

The narrowly scoped cancellation handling preserves other error logging and has focused regression coverage; tests were not executed.

Review effort: Balanced
Findings: None

What changed in this PR

Prevents expected health-monitor shutdown cancellation from producing error logs, addressing the rolling-upgrade flake tracked in #11408.

Changes:

• Handles cancellation after monitor shutdown without suppressing other exception types.
• Adds regression tests for shutdown cancellation and unexpected failures before and after stopping.

File	Description
test/​Orleans.Core.Tests/​Membership/​SiloHealthMonitorTests.cs	Adds three focused cancellation and error-logging tests.
src/​Orleans.Runtime/​MembershipService/​SiloHealthMonitor.cs	Filters expected shutdown cancellation in the monitoring loop.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[github-actions]

Code coverage

Metric	Pull request
Lines	83.28% (117,894 / 141,566)
Branches	72.76% (34,936 / 48,016)

Report-only conclusion: current-main baseline stale.

The newest successful coverage run tested 961a9bc, not current main 3d64611.

Coverage combines every CI test matrix job, including providers, CodeGen, .NET 8/10, Linux, Windows, and macOS, using canonical physical source and branch identities.

The comparison remains report-only while normal line and branch variance is calibrated.

Coverage details

• Pull request commit: 143ea0e1e6
• View the pull request test run and coverage artifacts
• View the trusted coverage reporting run