System performance
The following Key Performance Indicators are collected, aggregated, and reported:
| System Type | Aggregation | Performance Resource Type | Source |
|---|---|---|---|
| Physical Server | Hourly, daily, and monthly | Total average CPU usage | CPULOAD |
| Total size and used space (kB) of every local filesystem, and accumulated over all local filesystems | FILESYSTEMS | ||
| Total amount of and used virtual memory | PAGINGSPACE | ||
| Amount of used physical memory | MEMORY | ||
| Network I/O transmitted and received kilobytes | NETWORK_IO custom check | ||
| Disk I/O number of reads and writes | DISK_IO custom check | ||
| Disk I/O read and written kilobytes | |||
| Network Response time | NW_RESPONSE custom check | ||
| HTTP Response time | HTTP_RESPONSE custom check | ||
| SAP Instance (ABAP and ABAP+Java) | Hourly, daily, and monthly | Maximum number of concurrent users | Monitoring Tree ActivitySnapshot ActiveUsersCount |
| Daily and monthly | Average dialog response time and number of dialog steps | Daily check | |
| Average response time and number of dialog steps for named transactions | |||
| Average response time and number of dialog steps of all named transactions together per instance | |||
| Top N transaction response times sorted by average response time, and sorted by number of steps | |||
| CCMS performance counters | CCMS and CCMS_MONSET custom checks | ||
| SAP Instance (Java-only) | Hourly, daily, and monthly | J2EE garbage collection runtime ratio | JVMGarbageCollector |
| J2EE DSR average action response time | Average of sum of times (CPU, gen, load, net...) per DSR record with action types unequal to 'unknown' | ||
| J2EE DSR number of actions | |||
| JVM memory used and allocated space | JVM_MEMORY | ||
| SAP Instance (Web Dispatcher) | Hourly, daily, and monthly | Number of connections | WD_ConnectionStat |
| Number of threads | WD_ThreadStat | ||
| Number of queues | WD_QueueStat | ||
| SAP System (ABAP and ABAP+Java) | Hourly, daily, and monthly | Number of concurrent users | TH_USER_LIST |
| Number of Transports | TMS_ImportStat | ||
| Daily and monthly | Average dialog response time and number of dialog steps | Daily check | |
| Average response time and number of dialog steps for named transactions | |||
| Average response time and number of dialog steps of all named transactions together per system | |||
| Top N transaction response times sorted by average response time, and sorted by number of steps | |||
| Hourly, daily, and monthly | End-to-End ABAP, Java, and HTTP response times | End to End Application Monitoring | |
| CCMS performance counters | CCMS and CCMS_MONSET custom checks | ||
| Database (All Database types, except MySQL and PostgreSQL) | Hourly, daily, and monthly | Database size and used space | ADA_DATAAREA, DB2_TABLESPACES, MSS_DBUSAGE, SYB_DataSpaces, HDB_Disks, ANY_DBSpaces, SIQ_DataSpaces and ORA_TABLESPACES |
| Database (MaxDB) | Hourly, daily, and monthly | Number of sessions | Count of sessions belonging to connected users |
| Data cache hit ratio | |||
| Catalog cache hit ratio | |||
| Database (Microsoft SQL Server) | Hourly, daily, and monthly | Number of logons | OS Performance value of counter with name 'User Connections' |
| Buffer cache hit ratio | |||
| Lock wait time | |||
| Number of deadlocks | |||
| Database (Oracle) | Hourly, daily, and monthly | Tablespace size and used space | ORA_TABLESPACES |
| Number of logons | 'logons current' from V$SYSSTAT | ||
| Data buffer hit ratio | |||
| DD cache hit ratio | |||
| User/recursive calls ratio | |||
| Sorts in memory | |||
| Short table scans | |||
| Number of waiting sessions | |||
| Enqueue deadlocks, timeouts, and waits | |||
| Database (DB2) | Hourly, daily, and monthly | Tablespace size and used space | DB2_TABLESPACES |
| Number of connections | Sum of local and remote connections SNAPDBM | ||
| Buffer pool hit ratio | |||
| Sort overflows % | |||
| LSN gab clean triggers | |||
| Time database waited on locks | |||
| Lock waits | |||
| Deadlocks | |||
| Lock escalations | |||
| Database (SAP HANA) | Hourly, daily, and monthly | Disk size/usage of named volume partition | Performance data tables M_VOLUMES and M_VOLUME_SIZES |
| Physical memory size/usage by host | |||
| Peak memory limit size/usage | |||
| CPU load by host | |||
| Number of total lock waits | |||
| Total lock wait time per server | |||
| Backup Size | |||
| Backup Throughput | |||
| Database (SAP ASE) | Hourly, daily, and monthly | Number of connections | Count of suids in master.sysprocesses |
| Cache hit ratio | |||
| Database (SAP SQL Anywhere) | Hourly, daily, and monthly | Number of connections | ConnCount from sa_db_properties |
| Cache hit ratio | |||
| Checkpoint urgency | |||
| Recovery urgency | |||
| Database (SAP IQ) | Hourly, daily, and monthly | Main Cache Size | SIQ_Cache |
| Temp Cache Size | |||
| Main Cache Used | |||
| Temp Cache Used | |||
| Main Cache Hit Ratio | |||
| Temp Cache Hit Ratio | |||
| Catalog Cache Max Size | |||
| Catalog Cache Used | |||
| Catalog Cache Hit Ratio | |||
| Memory Allocated | |||
| Connections Active | SIQ_Connections |
Performance Indicators are shown in the UI and reported in the Service Level Reports for those Systems where Performance Data Collection is turned on. This can be turned on or off on a per-System (via PerfDataCollection Monitoring Parameter) level.
Performance Indicators are aggregated automatically into hourly, daily, and monthly values, whereas hourly and daily values are kept for six months and monthly values are kept for three years in Avantra (see also Performance.PERFDATA_KEEPTIME_MONTHLY).
Configuring Performance Data Collection
Most of the configuration options have already been described at Performance Data Collection, but we summarize them here and describe some more options:
- The PerfDataCollection Monitoring Parameters define if or if not Performance Data Collection is executed at all. This parameter is available for Servers, Databases, SAP Instances, and SAP Systems.
- The Avantra Server properties Performance.PERFDATA_KEEPTIME and Performance.PERFDATA_KEEPTIME_MONTHLY define how long Performance data is kept within Avantra.
- If you want to collect response times of named transactions, you have to define them using the PerfDataTransactions Monitoring Parameter.
- You can configure the Top N response time statistics using PerfDataTARuntimeStepsMinPct, PerfDataTARuntimeTop, PerfDataTAStepsTop, and PerfDataTopNIncludeReports Monitoring Parameters.
- If the Checks the Performance Data Collection is based upon are deactivated, no Performance Data is collected.
- Some of the thresholds configuring the underlying Checks configure also the display of the Performance Data. If Warning or Critical thresholds are exceeded the points in the Performance Data histograms are colored accordingly. These are:
- DialogRespTimeWarn and DialogRespTimeCrit for the Transaction Response Times
- JavaASAvgRespTimeWarn and JavaASAvgRespTimeCrit for the J2EE Response Times
- End-to-End Response Times are only available if you have configured End Points.
Displaying Performance Data
You can display performance data for every system in Avantra. Choose Systems > Servers, Systems > SAP Instances, Systems > SAP Systems, or Systems > Databases and proceed to the Performance tab.
There is a list or even a tree of charts on the left-hand side. You can double-click an item to add a chart to the page and click X to remove it.
You may select the time range of the data globally for all charts displayed. There are pre-defined ranges (Last 24 Hours, Last Week, Last Month, Last Year), or you fill in arbitrary values for Since and Until.
In addition, you have the option to drill into every single chart by dragging a rectangle around the desired region of the chart with your mouse. Click on the arrow icon to restore the original size.
The overall size of the charts may be changed with + and -.
For every chart there is an option to display raw data in a table, and to export it in a CSV file (using the file icon). Once you have arranged all charts on the page, you may export them into a PDF file.
Charts are scaled automatically based on the maximum and minimum values of the data. This may be misleading, especially when it comes to percentage values. There may be situations where the chart draws awfully high (visual) maximum values while the y-Axis in fact shows a range between, for example, 0.3% and 0.6%. In this case, it may be more reasonable to adjust the range of the y-Axis to the values between 0 and 100%. You can accomplish this using the arrow icon.
As mentioned earlier, there may be a tree of charts on the left-hand side. For example, for an SAP System, you can display all the SAP System-related Performance data, but you can also include the ones of the Database instance, other SAP Instances, or even the ones of the Servers comprising the SAP System. This allows correlating, for example, Transaction Response times with CPU values or Database locks.
You need the View Systems Permission in order to view Performance data.
Configuring Custom Performance Data Resource Types
Custom performance data collection can be done via the RUN_PROG Custom Check, or the API Documentation of the Custom Checks SQL_QUERY and RUN_JS. But prior to this, you need to define your own custom performance resource types. To do this, follow the steps below:
- Select Configuration > Performance Data.
- Click New to create a new Performance resource type. A name for the new Performance resource type is required and it has to be defined for which System Type it's relevant. If the System Type is either SAP Instance or SAP System, additionally, the Instance Type can be selected. The checkbox Create default chart definition is checked by default. It means along with a new Performance resource type a chart definition will be created.
- Click Finish, and the new Performance resource type is ready and can be used to collect custom performance data.
Displaying Custom Performance Data
To display the custom performance data, it's necessary to create a custom chart definition. When the performance resource type is created, it's possible to create a chart definition as well. But the chart definition can be created afterwards or a second chart definition can be created.
- Select Configuration > Performance Data.
- Switch to the Charts.
- Open a chart definition or create a new one to modify colors, axis labels, etc.