@AlessandroPic1

usually no big increase is needed if you’re not adding Constellation, Cassandra, or new features, but plan a buffer. Here’s the thing: move your containers to Pega 24.2-supported JDK (17) and base image, and keep the same CPU/memory to start; raise JVM heap by ~10–20% only if GC pauses grow under your normal load. Make sure Search (SRS) nodes can handle a full reindex after upgrade; give them short-term extra CPU and I/O during reindex only, then scale back. Queue processors and batch jobs changed a lot since 8.8, so watch their throughput and raise worker threads or node CPU before adding RAM. On the DB, you generally keep the same vCPU/RAM, but you need extra temporary disk/IO for schema upgrade, reindex, and statistics refresh; plan a maintenance window with 20–30% IOPS headroom. After go-live, compare AWR/pg_stat/latency against your 8.8 baseline and tune caches (rule/data) rather than throwing hardware. If you use BIX/DFs, test peak runs—these are the first to expose resource gaps. Final tip: load test your top 5 journeys at 1.2–1.5× expected traffic on 24.2 and only then right-size.