fix(ci): match E2E app tags literally, make them unique per run, and never stop the caller - #8721

Summary

• stop-session.sh matches E2E_APP=<tag> as a fixed string (grep -Flzx). A tag like s.im, or a prefix like sc, no longer matches another app's E2E_APP=scim.
• Each HTTP E2E app's tag is now <app>-<run id>-<attempt>-<shell pid>, so it is unique among the runner user's processes.
• The script never signals itself, its ancestors, or its own subshells, even when the caller exported the same tag. Before, it SIGTERMed itself and its parent shell (rc=143). It reads /proc/<pid>/stat without forking, so scans stay cheap.
• A surviving Next telemetry flush is labelled "(expected telemetry flush)", so an unexpected escapee stands out.

Type of Change

• Bug fix

Testing

• Ran the checks in a node:22-bookworm container, and all passed:
  • a s.im tag and a sc prefix tag each stop only their own app and leave E2E_APP=scim running;
  • with the caller exporting the same tag, the caller survives and the script exits 0, while the app and a tagged escapee in its own session are both stopped;
  • the flush label is printed.
• On Linux, every earlier session check still passes:
  • a member in its own process group is stopped;
  • a member that ignores SIGTERM is SIGKILLed at 10s;
  • a process spawned during shutdown is stopped;
  • a detached tagged writer is stopped;
  • a leader that ignores SIGTERM is SIGKILLed at 10s.
• Stop-then-wipe loop on Linux: 6 of 6 rounds had no flush left and nothing written under .next afterwards.
• Full HTTP E2E app sequence on Linux (8 cores, cold .next per app): 3 of 3 passed.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)