@Iñaki Ruiz Install HFIX-C681 only if you see daily web node instability after upgrading to 23.1.4—examples: repeated node unresponsive alerts, rising JVM heap/thread count, frequent 5xx from web tier, or manual restarts to recover. First, confirm it in lower environments: capture 48–72 hours of baseline metrics (GC logs, heap usage, thread/requestor counts, health/ALERT logs) and note any patterns leading to restarts. Review the hotfix README from Pega Support; if the affected engine/UI components or jar versions match your stack, proceed. Take a full backup, deploy the hotfix on a single non-prod node, and clear caches/restart. Post-install, run smoke tests (login, portals, case open/submit, REST calls) and then soak for 72 hours while monitoring GC, CPU, requestor pools, and PEGA0001/PEGA0020 alerts; success = no forced restarts and stabilized memory/threads. If stable, roll to all web nodes in prod during a maintenance window. If symptoms persist or you don’t see the pre-fix signatures, open an SR with your logs; the issue might be configuration (e.g., DSS, custom auth filters, load-balancer timeouts) rather than this hotfix.