System Stat Summary Interval (PI)

Function

The System Stat Summary Interval (PI) record stores performance data, taken at specific intervals since the start of an instance, about key performance indicators.

Default and changeable values

ItemDefault valueChangeable
Collection Interval60Y
Collection Offset0Y
LogYesY
LOGIF(Blank)Y

ODBC key fields

None

Lifetime

From the creation to the deletion of an Oracle instance

Record size

Fields

PFM - View name
(PFM - Manager name)
DescriptionSummaryFormatDeltaSupported versionData source
Block Changes/Tran
(BLOCK_CHANGES_PER_TRANSACTION)
Rate at which each transaction executed DML operation #2AVGdoubleYes#8Alldb block changes / user commits
Block Get/sec
(BLOCK_GET_RATE)
Rate at which the application referenced the database#2 AVGdoubleYes#8All(db block gets + consistent gets) / second in interval
Block Visits/Tran
(BLOCK_VISITS_PER_TRANSACTION)
Number of times a work database was loaded per transaction#2AVGdoubleYes#8All(db block gets + consistent gets) / user commits
Buffer Busy Wait %
(BUFFER_BUSY_WAIT_PERCENTAGE)
Percentage ratio of buffer busy waits#2AVGdoubleYes#8All(V$SYSTEM_EVENT.TOTAL_WAITS where EVENT = 'buffer busy waits' / (consistent gets + db block gets)) * 100
Cache Hit %
(CACHE_HIT_PERCENTAGE)
Buffer cache usage#2AVGdoubleYes#8All(1 - (physical reads cache / (consistent gets from cache + db block gets from cache))) * 100
Call/sec
(CALL_RATE)
Work demand rate applied to the instance by all work resources#2AVGdoubleYes#8Allrecursive calls + user calls / seconds in interval
Calls/Tran
(CALLS_PER_TRANSACTION)
Rate at which client requests were executed per transaction. You can use the value of this field to detect a change in the application or the utilization status. This value may increase significantly if unique queries increase.#2AVGdoubleNoAlluser calls / user commits
Changed Block %
(CHANGED_BLOCK_PERCENTAGE)
Percentage ratio of difference between query and database manipulation language (DML) in the database application. This value changes according to indexes and application utilization status.#2AVGdoubleYes#8All(db block changes / (block gets + consistent gets)) * 100
Consistent Change %
(CONSISTENT_CHANGE_PERCENTAGE)
Percentage indicating the extents for which the application needed to check the consistency of read operations#2AVGdoubleYes#8All(consistent changes / consistent gets) * 100
Continued Row %
(CONTINUED_ROW_PERCENTAGE)
Percentage ratio of continued rows. The value is close to 0 unless the application handles LONG columns. #2AVGdoubleYes#8All(table fetch continued row / (table fetch by rowid + table scan rows gotten)) * 100
Current Logons
(CURRENT_LOGONS)
Number of current logons to Oracle Database#2AVGlongNoAllV$SYSSTAT.VALUE
Deadlocks
(LOCK_DEADLOCKS)
Number of deadlocks caused by locked DML processing#2AVGdoubleYesAllV$SYSSTAT.VALUE
Dict Cache Get Miss %
(DICTIONARY_CACHE_GET_MISSES_PERCENTAGE)
Percentage ratio of data requests issued due to cache miss#2AVGdoubleYes#8All(SUM(V$ROWCACHE.GETMISSES) / SUM(V$ROWCACHE.GETS)) * 100
Disk Sorts
(SORTS_DISK)
Number of disk sort operations#2AVGdoubleYesAllV$SYSSTAT.VALUE
Free List Wait Events
(FREE_LIST_WAIT_EVENTS)
Number of wait events in the free list#2AVGdoubleYesAllV$WAITSTAT.COUNT where class = 'free list'
I/O Ops/sec
(IO_RATE)
Number of I/O operations per second#2AVGdoubleNoAll(physical reads + physical writes) / seconds in interval
Lib Cache Miss %
(LIBRARY_CACHE_MISS_PERCENTAGE)
Library cache miss rate. This field means the ratio of times the allocated objects in library cache are reloaded. As the value of this field increases, the amount of resources in use also increases. #2AVGdoubleYes#8All(SUM(V$LIBRARYCACHE.RELOADS) SUM(V$LIBRARYCACHE.PINS)) * 100
Lock Conversions
(LOCK_CONVERSIONS)
Number of enqueues (locks) whose mode was changed (such as from share to lock)#2AVGdoubleYesAllV$SYSSTAT.VALUE
Lock Hit %
(LOCK_HIT_PERCENTAGE)
Percentage ratio of the number of accesses to data blocks to the total number of data block accesses (limited to accesses that do not involve lock conversion; also valid only for parallel servers)#2AVGdoubleYes#8All((consistent gets - global lock converts (async)) / consistent gets) * 100
Lock Releases
(LOCK_RELEASES)
Number of times enqueuing (locking) was released#2AVGdoubleYesAllV$SYSSTAT.VALUE
Lock Requests
(LOCK_REQUESTS)
Number of times enqueuing (locking) was requested#2AVGdoubleYesAllV$SYSSTAT.VALUE
Lock Timeouts
(LOCK_TIMEOUTS)
Number of times an enqueuing (locking) request was not permitted within the allocated time#2AVGdoubleYesAllV$SYSSTAT.VALUE
Lock Waits
(LOCK_WAITS)
Number of times an enqueuing (locking) request was placed in wait status. The difference between the enqueuing requests count and the enqueuing waits count is the number of times the request was not treated as an enqueuing request. #2AVGdoubleYesAllV$SYSSTAT.VALUE
Logical Reads
(LOGICAL_READS)
Number of logical read operations in read consistency mode and number of requests to the current copy of blocks#2AVGdoubleYesAlldb block gets + consistent gets
Memory Sorts
(SORTS_MEMORY)
Number of sort operations in memory#2AVGdoubleYesAllV$SYSSTAT.VALUE
Non-Index Lookups %
(NON_INDEX_LOOKUPS)
Percentage ratio of full table scans that do not involve caching#2AVGdoubleYes#8All(table scans (long tables) / (table scans (short tables) + table scans (long tables))) * 100
Physical Reads
(PHYSICAL_READS)
Number of physical read operations on database block from disk#2AVGdoubleYesAllV$SYSSTAT.VALUE
Physical Writes
(PHYSICAL_WRITES)
Number of physical write operations on the disk by DBWR#2AVGdoubleYesAllV$SYSSTAT.VALUE
Ping Write %
(PING_WRITE_PERCENTAGE)
A large value means that lock conversion occurred frequently#2AVGdoubleYes#8All(DBWR cross instance writes / physical writes) * 100
Read/sec
(READ_RATE)
Number of read operations per second#2AVGdoubleNoAllphysical reads / seconds in interval
Record Time
(RECORD_TIME)
Collection termination time for the performance data stored in the record#1COPYtime_tNoAllRemote Monitor Collector
Record Type
(INPUT_RECORD_TYPE)
Record name (always PI)#1COPYstring(4)NoAllRemote Monitor Collector
Recursive Calls
(RECURSIVE_CALLS)
Number of user calls processed#2AVGdoubleYesAllV$SYSSTAT.VALUE
Recursive To User Call %
(RECURSIVE_TO_USER_CALL_PERCENTAGE)
Correct values cannot be collected in this field.
Percentage indicating overhead#2
AVGdoubleYes#8All(recursive calls / user calls) * 100
Redo Alloc Immediate %
(REDO_ALLOC_IMMEDIATE_HIT_PERCENTAGE)
Success rate of immediately acquiring REDO allocation latch#2AVGdoubleYes#8All(1 - (V$LATCH.IMMEDIATE_MISSES / (V$LATCH.IMMEDIATE_GETS + V$LATCH.IMMEDIATE_MISSES))) * 100 where V$LATCH.NAME = 'redo allocation'
Redo Alloc Willing to Wait %
(REDO_ALLOC_WILLING_TO_WAIT_HIT_PERCENTAGE)
Success rate of acquiring REDO allocation latch from cache#2AVGdoubleYes#8All(1 - (V$LATCH.MISSES / V$LATCH.GETS)) * 100 where name = 'redo allocation'
Redo Copy Immediate %
(REDO_COPY_IMMEDIATE_HIT_PERCENTAGE)
Success rate of immediately acquiring REDO copy latch#2AVGdoubleYes#8All(1 - (V$LATCH.IMMEDIATE_ MISSES / (V$LATCH.IMMEDIATE_GETS + V$LATCH.IMMEDIATE_MISSES))) * 100 where V$LATCH.NAME = 'redo copy'
Redo Copy Willing to Wait %
(REDO_COPY_WILLING_TO_WAIT_HIT_PERCENTAGE)
Success rate of acquiring REDO copy latch from cache#2AVGdoubleYes#8All(1 - (V$LATCH.MISSES / V$LATCH.GETS)) * 100 where name = 'redo copy'
Redo Log Space Requests
(REDO_LOG_SPACE_REQUESTS)
Number of times Oracle must wait for disk spaces to be allocated to REDO log entry because the active log file is full. #2AVGdoubleYesAllV$SYSSTAT.VALUE
Redo Log Space Wait %
(REDO_LOG_SPACE_WAIT_PERCENTAGE)
Whether or not memory allocation is appropriate#2AVGdoubleYes#8All(redo log space requests / redo entries) * 100
Row Source %
(ROW_SOURCE_PERCENTAGE)
Percentage ratio of rows obtained by full-table scan#2AVGdoubleYes#8All(table scan rows gotten / (table fetch by rowid + table scan rows gotten)) * 100
SQL Executing
(SQL_EXECUTING)
This field is not supported.
Number of current SQL executions
AVGdoubleNoNot supportedRemote Monitor Collector
SQL Net Bytes Rcvd
(SQL_NET_BYTES_RECEIVED)
Amount of data received from clients via SQL*Net#2AVGdoubleYesAllV$SYSSTAT.VALUE
SQL Net Bytes Sent
(SQL_NET_BYTES_SENT)
Amount of data sent to clients via SQL*Net#2AVGdoubleYesAllV$SYSSTAT.VALUE
Session CPU Usage
(SESSION_CPU_USAGE)
If the value of the TIMED_STATISTICS parameter in the init.ora file is TRUE, the CPU time used (in 1/100 seconds); if the parameter value is FALSE, this field contains 0.#2AVGdoubleYesAllV$SYSSTAT.VALUE
Session Cursor Cache Count
(SESSION_CURSOR_CACHE_COUNT)
Number of analysis calls for the session's cursor calls#2AVGdoubleYesAllV$SYSSTAT.VALUE
Session Cursor Cache Hit %
(SESSION_CURSOR_CACHE_HIT_PERCENTAGE)
Hit rate of analysis calls in session's cursor cache#2AVGdoubleNoAll(session cursor cache hits / session cursor cache count) * 100
Session Cursor Cache Hits
(SESSION_CURSOR_CACHE_HITS)
Number of times a cursor was found in the session's cursor cache by analysis call#2AVGdoubleYesAllV$SYSSTAT.VALUE
Session PGA Memory
(SESSION_PGA_MEMORY)
Size of PGA memory currently being used in bytes#2AVGdoubleNoAllV$SYSSTAT.VALUE
Session UGA Memory
(SESSION_UGA_MEMORY)
Size of used session memory in bytes#2AVGdoubleNoAllV$SYSSTAT.VALUE
Sort Overflow %
(SORT_OVERFLOW_PERCENTAGE)
Percentage ratio of sort operations using a temporary segment #2AVGdoubleYes#8All(sorts (disk) / (sorts (memory) + sorts (disk))) * 100
Start Time
(START_TIME)
Collection start time for the performance data stored in the record#1COPYtime_tNoAllRemote Monitor Collector
Sys Undo Blk Wait Events
(SYSTEM_UNDO_BLOCK_WAIT_EVENTS)
Number of block wait events for the system's rollback segment #2AVGdoubleYesAllV$WAITSTAT.COUNT where class = 'system undo block'
Sys Undo Hdr Wait Events
(SYSTEM_UNDO_HEADER_WAIT_EVENTS)
Number of header block wait events for the system's rollback segment #2AVGdoubleYesAllV$WAITSTAT.COUNT where class = 'system undo header'
Total Logons
(TOTAL_LOGONS)
Number of logins to the server #2AVGdoubleYesAllV$SYSSTAT.VALUE
Total SQL Executions
(TOTAL_SQL_EXECUTIONS)
Total number of SQL executions #2AVGdoubleYesAllexecute count
Trans/sec
(TRANSACTION_RATE)
Number of transactions per second#2AVGdoubleNoAlluser commits / seconds in interval
Undo Blk Wait Events
(UNDO_BLOCK_WAIT_EVENTS)
Number of block wait events for other rollback segments#2AVGdoubleYesAllV$WAITSTAT.COUNT where class = 'undo block'
Undo Hdr Wait Events
(UNDO_HEADER_WAIT_EVENTS)
Number of header block wait events for other rollback segments#2AVGdoubleYesAllV$WAITSTAT.COUNT where class = 'undo header'
User Calls
(USER_CALLS)
Number of user calls processed#2AVGdoubleYesAllV$SYSSTAT.VALUE
User Calls / Parse
(USER_CALLS_PER_PARSE)
Application management status in the context area #2AVGdoubleYes#8Alluser calls / parse count (total)
User Commits
(USER_COMMITS)
Number of transactions#2AVGdoubleYesAllV$SYSSTAT.VALUE
User Rollback %
(USER_ROLLBACK_PERCENTAGE)
Failure rate of application transactions #2AVGdoubleNoAll(user rollbacks / (user commits + user rollbacks)) * 100
User Rollbacks
(USER_ROLLBACKS)
Number of rollbacks#2AVGdoubleYesAllV$SYSSTAT.VALUE
VA DeviceID
(VADEVICEID)
Device ID of virtual agent.--string(256)NoAllRemote Monitor Collector
Write %
(WRITE_PERCENTAGE)
Percentage ratio of write operations #2AVGdoubleNoAll(physical writes / (physical reads + physical writes)) * 100
Writes/sec
(WRITES_RATE)
Number of write operations per second#2AVGdoubleNoAllphysical writes / seconds in interval
#8
The field value is calculated based on the delta of the amount of collected data from Oracle Database.