Use case · SAP HANA
SAP HANA monitoring, with a user that only reads
SUPtime queries HANA over SQL with a dedicated read-only user, and puts memory, slow queries, backups and disks next to the Basis probes of the same system.
What it reads
The system views, one by one
Every value has a precise source. Without database credentials, the tab says so instead of showing numbers.
| Source | What it shows |
|---|---|
| M_SERVICE_MEMORY | Memory used by the HANA services |
| M_HOST_RESOURCE_UTILIZATION | Host CPU and memory |
| M_SQL_PLAN_CACHE | The slowest queries |
| M_TABLES · M_TABLE_PERSISTENCE_STATISTICS | The largest tables |
| M_BACKUP_CATALOG | Outcome and age of the last backup |
| M_DISKS | Data and log disk usage: when they fill up, HANA stops |
| M_BLOCKED_TRANSACTIONS · M_TRANSACTIONS | Blocked transactions and who blocks them |
| _SYS_STATISTICS.STATISTICS_CURRENT_ALERTS | Statistics server alerts, with SAP's own explanation and recommended action |
| M_SYSTEM_REPLICATION | System replication status: active, in error, or not configured |
| M_LICENSES | Memory limit and license expiry |
| M_DELTA_MERGE_STATISTICS · M_LOG_SEGMENTS · M_SAVEPOINTS | Failed delta merges, unfreed log segments, savepoint critical-phase duration |
| M_TABLE_PARTITIONS · M_OUT_OF_MEMORY_EVENTS · M_CS_UNLOADS | Tables near the row limit per partition, out-of-memory events, unloads caused by low memory |
| SYS.USERS (own user) | Whether the monitoring user is still active, its roles, when it expires |
| M_XS_APPLICATIONS · M_XS_APPLICATION_ISSUES | XS Classic applications and activation errors |
| M_XS_SESSIONS · _SYS_XS.JOB_LOG | XS Classic sessions and jobs |
Every morning
The questions a Basis admin asks about HANA
Did last night's backup run?
Outcome and age of the last backup from the catalog: an old backup shows up before you need it.
Are the disks holding?
Data and log disk usage. A full volume stops the database: better to know with room to spare.
Who is using the memory?
Memory per service and host resources, read at every poll.
What is slowing the system down?
Slowest queries from the plan cache, largest tables and blocked transactions.
HANA health
Not just load: the diagnostics HANA already computes
HANA's statistics server already evaluates dozens of conditions. SUPtime collects them instead of duplicating them, with the standard MONITORING role.
Statistics server alerts
With rating, explanation and SAP's own recommended action: not just a number.
System replication
Active, in error, or "not configured": never a guess when the view comes back empty.
Delta merge, log, savepoints
The three early signals of a database about to slow down or lock up, before a user notices.
The two-billion wall
A column store partition stops at 2^31 rows. Tables approaching the limit show up months ahead, alongside out-of-memory events and unloads forced by low memory.
The monitoring user itself
Whether it is still active, when it expires, which roles it has: a locked-out user shows up on the tab, not in the logs.
XS runtime
XS Classic monitored, XS Advanced declared
Two platforms with the same short name and nothing in common: the tab always says which one it is looking at.
XS Classic
Installed applications, activation errors, sessions and runtime jobs.
XS Advanced
Detected from the active services and declared not supported, instead of showing zero applications.
XS jobs
They need a dedicated grant. Without it, the failed-jobs counter shows "—", not "0".
Access
A read-only user, not the schema owner
Never SAPHANADB
The owner of the application schema can write to every table. SUPtime only reads, so it uses a dedicated user.
Standard role, not custom
HANA's predefined monitoring role, plus catalog read: nothing beyond what is required.
Script and checks
The documentation includes the creation script and the queries to verify the privileges.
One system, two channels
HANA and ABAP in the same view
Basis probes over RFC
Dumps, jobs, locks and the Security Audit Log of the ABAP stack running on that database.
Two SIDs in the emails
The ABAP SID and the database SID do not always match: alerts carry both.
Rules on any monitor
For one system, a landscape or all of them, with recovery notifications.
Declared limits
What it does not cover
Knowing what a monitor does not see matters as much as knowing what it sees.
- This page covers HANA: Oracle has its own use case. Other databases (ASE, DB2, SQL Server) are not supported.
- XS Advanced: detected, not monitored.
- System replication: only the status it reads (active, in error, not configured). No failover or cluster management.
- Features that would need more than the MONITORING role: not built, as a security choice.
- Forecasts: SUPtime shows the values it read, it does not estimate future ones.
Want to see it on your HANA?
We will show you the tabs on a real system and the script for the read-only user.