Understanding SAP HANA Memory: A Beginner’s Guide

SAP HANA is an in-memory database, so memory isn’t just a resource — it’s the database. Understanding how it’s used is the first step to keeping a system stable.

The key numbers

• Physical memory — the RAM installed on the host.
• Global allocation limit — the maximum memory HANA is allowed to use.
• Used memory — what HANA is actually using right now, including data, working space and code.

When used memory approaches the allocation limit, HANA starts unloading tables and, eventually, queries fail with out-of-memory errors.

Where to look

SELECT HOST,
       ROUND(INSTANCE_TOTAL_MEMORY_USED_SIZE / 1024 / 1024 / 1024, 1) AS USED_GB,
       ROUND(ALLOCATION_LIMIT / 1024 / 1024 / 1024, 1)                AS LIMIT_GB
FROM   M_HOST_RESOURCE_UTILIZATION;

Run this regularly and trend it. A steady climb tells you far more than a single snapshot.

Common causes of memory pressure

1. Large tables that could be partitioned or archived.
2. Expensive queries creating big intermediate results.
3. Growth that simply outpaced the original sizing.

Set alerts well below the limit — 80% is a sensible starting point — so you have time to act.