fix(client-runtime): tolerate transient foreground probe timeouts

[Zeus-Deus]

Problem

When the host is temporarily resource constrained, a foreground health probe can time out even though the existing session is still valid. Immediately replacing that session can start a reconnect while the same host is still stalled, leaving the client unusable until it recovers.

Fix

Keep the existing desktop/web session after the first foreground probe timeout. A successful probe resets the count, while a second consecutive timeout still triggers normal recovery. Definite probe failures and mobile resume behavior are unchanged.

Testing

• Focused connection supervisor suite: 35 tests passed
• Client-runtime typecheck
• Targeted lint and formatting checks

Model: GPT-5; harness: Codex.

---

Note

Medium Risk
Changes live connection monitoring and reconnect timing in client-runtime; behavior is well-covered by supervisor tests but affects when sessions are replaced under load.

Overview
Foreground health probes on application-active no longer tear down the session on a single 15s timeout when the host is briefly stalled. The supervisor keeps the existing lease, logs a warning, and only triggers normal reconnect recovery after two consecutive timeouts (FOREGROUND_PROBE_TIMEOUTS_BEFORE_RECONNECT).

A successful probe resets the counter. Definite probe failures and application-active-probe (mobile, 3s timeout) behavior are unchanged.

Tests cover: one timeout stays connected; two timeouts → backoff/release; success between timeouts resets the count.

^{Reviewed by Cursor Bugbot for commit 3fdbd15. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Tolerate transient foreground liveness probe timeouts in EnvironmentSupervisor

• During application-active wakeups, a single probe timeout no longer triggers a reconnect; the supervisor only reconnects after 2 consecutive timeouts (FOREGROUND_PROBE_TIMEOUTS_BEFORE_RECONNECT = 2).
• A successful probe resets the timeout counter to 0, so isolated timeouts are always tolerated.
• application-active-probe signals retain the existing fail-fast behavior on timeout.
• Two new tests cover consecutive timeout reconnection and counter reset after a successful probe.

^{Macroscope summarized 3fdbd15.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 61f1d7d2-365c-4fe9-a512-c2ef24bb1bdd

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR changes runtime behavior for connection supervision - tolerating transient foreground probe timeouts before reconnecting instead of failing immediately. This modifies core connection handling logic and should be reviewed by someone familiar with the supervisor implementation.

^{You can customize Macroscope's approvability policy. Learn more.}