Key Performance Indicators and Key Performance Areas for Replication

Lists and describes the key performance indicators (KPIs) and key performance areas (KPAs) that provide the statistics displayed on Replication screens and charts in SAP Control Center.

Each Replication KPA includes a subset of the KPIs.

Key Performance Indicators
KPI Description Data Collection
Managed Object States
Primary State Managed object primary state. Monitoring Statistics
Secondary State Managed object seconday state.
Replication Server Performance Statistics
Insert commands Insert command on active object. Performance Counters
Update commands total Update command on active object.
Delete commands total Delete command on active object.
Writetext commands total Writetext command on active object.
Execute commands total Execute command on active object.
AIO collect time total Amount of time that the Asynchronous I/O Daemon spent in collecting async I/O requests in which Daemon is obliged to poll until notified of their completion.
AIO after poll total Number of pending asynchronous I/O requests just after the Async I/O Daemon finishes polling for their completion.
AIO before poll total Number of pending asynchronous I/O requests just before the Async I/O Daemon goes off to poll for their completion.
AIO poll time total Amount of time that the Asynchronous I/O Daemon spent in polling I/O completions. This includes the time taken to notify AIO clients of their I/O requests completion.
AIO read retry total The number of times an async read is made and the OpSys tells the client that it must try again.
AIO write retry total The number of times an async write is made and the OpSys tells the client that it must try again.
Alarm milliseconds spent total Number of milliseconds actually spent before the alarm was expired.
Alarm cancel total Number of alarms cancelled.
Alarm examined total Number of examined alarms to determine whether they are expired or not.
Alarm expired cancelled total Number of alarms that were found canceled upon each 'handle expiration' execution.
Alarm expired total Number of alarms that have expired upon each 'handle expiration' execution.
Alarm milliseconds expired interval total Number of milliseconds until the next alarm expiration.
Alarm expired time total Amount of time elapsed while handling expired alarms.
Alarm milliseconds requested total Number of milliseconds originally requested for the alarm when it was set.
Alarm set total Number of alarms set.
Thread start total Total number of started threads.
Thread stop total Number of threads stopped.
Begins total Begin Tran statements sent to Replication Server System Database (RSSD).
Deletes total Delete statements sent to RSSD.
Inserts total Insert statements sent to RSSD.
Selects total Select statements sent to RSSD.
Cache exceed total Number of time System Table Services (STS) cached was exceeded.
Updates total Update statements sent to RSSD.
Replication Server
Replication Server State The state of the Replication Server. Monitoring Statistics
% Memory usage The percent of the current memory used by the server.
Maximum Latency The highest latency of any replication path in the server.
% Partition Usage The percent full of all partitions in the Replication Server.
Path
Path Status The state of the replication path between a primary and a replicate database. Monitoring Statistics
Path Latency The time between when a transaction is committed at the primary database and when it is committed at the replicate database.
Change in Path Activity The number of commands sent to the replicate database by the Data Server Interface (DSI) thread since the last monitoring cycle.
Path Data Loss Indicates possible data loss between the primary and replicate databases.
rs_ticket Count The ticket's unique sequence number.
Primary Database Start Time The time, in milliseconds, the rs_ticket was executed at the primary database.
Executor Thread Latency The number of milliseconds from the start, to the time the ticket passes through the executor thread.
Distributor Thread Latency The number of milliseconds from the primary Replication Server time, to the time the ticket passes through the distributor thread.
Replication Server Interface Latency The number of milliseconds from the distributor (DIST) thread time, to the time the ticket passes through the Replication Server Interface (RSI) thread.
Data Server Interface Thread The number of milliseconds from the RSI thread time, to the time the ticket passes through the DSI thread.
Path RDB Latency The number of milliseconds from the DSI thread time, to the time the ticket arrived at the replicate database.
Connection
Time closing an OB fadeout total The amount of time spent closing a connection marked to be faded out. Performance Counters
Time waiting on OB fadeout total The amount of time spent waiting for a connection marked to be faded out.
Commands read total Commands read from an inbound queue by a DIST thread.
Resolved delete statements total Delete commands encountered by a DIST thread and resolved by subscription resolution engine (SRE).
Unresolved statements total DIST commands with no subscription resolution that a DIST thread discarded. This implies either there is no subscription or the 'where' clause associated with the subscription does not result in row qualification.
Resolved insert statements total Insert commands that a DIST thread encountered, which SRE resolved.
Resolved update statements total Update commands that a DIST thread encountered, which SRE resolved.
TD Begins total Begin transaction commands that a DIST thread propagated.
TD Commits or Rollbacks total Commit or Rollback commands that a DIST thread processed.
Transactions processed total Transactions read by DIST thread from an inbound queue.
All large threads used total This counter is incremented each time a Large Parallel Transaction must wait because there are no available parallel DSI threads.
All threads used total This counter is incremented each time a Parallel Transaction must wait because there are no available parallel DSI threads.
Commands applied successfully total Commands that a DSI successfully applied to the target database.
Time dispatching large groups total Time spent by the Data Server Interface, scheduler thread (DSI/S) dispatching a large transaction group to a DSI/E. This includes the time spent finding a large group to dispatch.
Time dispatching regular groups total Time that the DSI/S spent dispatching a regular transaction group to a DSI/E.
Time finding regular groups total Time that the DSI/S spent finding a group to dispatch.
Time loading SQT cache total Time that the DSI/S spent loading Stable Queue Transaction (SQT) cache.
Time executing special group total Time that the DSI/S spent determining if a transaction is special, and executing it if it is.
DSI/E in sleep mode total Number of DSI/E threads that the DSI/S put to sleep prior to loading SQT cache. These DSI/E threads have just completed their transaction.
Time putting DSI/E in sleep mode total Time that the DSI/S spent putting free DSI/E threads to sleep.
Transaction groups read total Transaction groups that the DSI read. If grouping is disabled, grouped and ungrouped transaction counts are the same.
Ungrouped transactions read total Ungrouped transactions that the DSI read. If grouping is disabled, grouped and ungrouped transaction counts are the same.
Transaction groups committed total Transactions that a DSI thread successfully committed.
Transaction groups sent total Transaction groups that the DSI thread sent to the target database. A transaction group can contain at most dsi_max_xacts_in_group transactions. This counter is incremented each time a 'begin' for a grouped transaction is executed.
Transaction groups success total Transaction groups that a DSI thread successfully applied to a target database. This includes transactions that were successfully committed or rolled back according to their final disposition.
Transaction groups committed total Transactions in groups sent by a DSI thread that committed successfully.
Transaction groups sent total Transactions contained in transaction groups that a DSI thread sent.
Closed groups maximum bytes total Transaction groups that a DSI thread closed due to the next transaction that caused the dsi_xact_group_size to exceed.
Closed groups next new origin total Transaction groups that a DSI closed due to no open group from the origin of the next transaction. For example, there is a new origin in the next transaction, or the schedule forced a flush of the current group from the origin leaving no open group from that origin.
Closed groups partitioning total Transaction groups that a DSI thread closed because of a Transaction Partitioning rule.
Closed groups maximum log transaction total Transaction groups that a DSI thread closed due to the next transaction that caused the dsi_max_xacts_in_group to exceed.
Groups wait partitioning total Transaction groups forced to wait for another group to complete (processed serially based on Transaction Partitioning rule).
True return of rs_dsi_check_thread_lock total Number of rs_dsi_check_thread_lock invocations returning true. The function determined the calling thread holds locks required by other threads. A rollback and retry occurred.
Delete statements read total rs_delete commands that a DSI executor (DSIEXEC) thread processed.
Batch size total Size, in bytes, of commands batches, which a DSI submitted.
Batch time total Time, in milliseconds, to process command batches, which a DSI submitted.
Bytes succeed total Bytes that a DSI/E successfully applied to the target database.
Commands read total Commands read from an outbound queue by a DSIEXEC thread.
Commands succeed total Commands that a DSI/E successfully applied to the target database.
Function string mapping time total Time, in milliseconds, to perform function string mapping on commands.
Input command count total Number of input commands in command batches that a DSI submitted.
Output command count total Number of output commands in command batches that a DSI submitted.
Result time total Time, in milliseconds, to process the results of commands batches that a DSI submitted.
Sequencing check on command batches time total Time, in milliseconds, to check the sequencing on command batches, which required some kind of synchronization such as wait_for_commit.
Sequencing check on commits time total Time, in milliseconds, to check the sequencing on commits.
Transaction time total Time, in milliseconds, for a DSI/E thread to process transactions. This includes function string mapping, sending and processing results. A transaction may span command batches.
Errors deadlock total Times that a DSI thread failed to apply a transaction due to deadlocks in the target database (ASE Error 1205).
Errors out of lock total Times that a DSI thread failed to apply a transaction due to no locks available in the target database (ASE Error 1204).
Writetext total rs_writetext commands that a DSIEXEC thread processed.
Insert total rs_insert commands that a DSIEXEC thread processed.
MemUsed group total Memory that a DSI/S thread consumed for transaction groups.
Send RPC time total Time, in milliseconds, spent in sending RPCs to the RDS.
Send time total Time, in milliseconds, spent in sending command buffers to the RDS.
System transaction read total Internal system transactions that a DSIEXEC thread processed.
Transaction groups scheduled total Transactions groups scheduled to a DSIEXEC thread.
Ungrouped transactions scheduled total Transactions in transaction groups scheduled to a DSIEXEC thread.
Updates read total rs_update commands that a DSIEXEC thread processed.
Bytes received total Bytes that a RepAgent thread received. This size includes the TDS header size when in 'passthru' mode.
Mini abort total mini-abort commands (in Adaptive Server, SAVEXACT records) that a RepAgent thread processed. mini-abort instructs Replication Server to rollback commands to a specific OQIQ value.
Commands received total Commands that a RepAgent thread received.
Write wait time total The amount of time the RepAgent spent waiting for the Stable Queue Manager (SQM) Writer thread to drain the number of outstanding write requests to get the number of outstanding bytes to be written under the threshold.
Yield time total The amount of time the RepAgent spent yielding the processor while handling Log Transfer Language (LTL) commands each time the processor was yielded.
Execution time total The amount of time, in milliseconds, RepAgent user thread is scheduled by Open Client/Server (OCS).
Receive time total The amount of time, in milliseconds, spent receiving network packets or language commands.
Update rs_locator total Updates to RSSD. rs_locater where type = 'e' executed by a RepAgent thread.
Replication Agent Thread State The state of the Replication Agent thread in the Replication Server that is associated with the connection. Monitoring Statistics
DSI State The state of the DSI thread in the Replication Server that is associated with the connection.
DSI Activity The number of commands that the DSI thread sent to the replicate database.
DSI Activity per Second The number of commands that the DSI thread sent to the replicate database per second.
Transaction Ignored The number of transactions that were ignored.
Transaction Skipped The number of transactions that were skipped.
Transaction Succeeded The number of transactions that were successfully committed to the replicate database.
Transaction Failed The number of transactions that were failed.
Transaction Retried The number of transactions that were retried.
DSI SQT Cache Percent Used The percent of DSI SQT cache size used.
Transactions removed per second The number of transactions removed from the DSI SQT cache per second.
SQT cache low boundry The smallest size to which DSI SQT cache is configured before transactions start being removed from the cache.
Transactions read per second The number of transactions that the DSI read per second.
Commands per Transaction The number of commands per transaction.
Transactions per Group The number of transactions per group.
Commands per Batch The number of commands per batch.
Logical Connection
Operation in Progress State The state of the operation in progress. Monitoring Statistics
Route
Route State The state of the route at both the origin and destination Replication Servers. Monitoring Statistics
Route Activity The number of bytes sent to the destination Replication Server.
Route Activity per Second The number of bytes sent to the destination Replication Server per second.
Alarm time spent total Number of milliseconds actually spent before an alarm expired. Performance Counters
Time spent cancelling alarms total Amount of time elapsed while canceling an alarm.
Expired alarms to be cancelled total Number of alarms that were found to have been expired while attempting to cancel them.
Alarm required time total Number of milliseconds originally requested for the alarm when it was set.
Alarm set total Amount of time elapsed while setting an alarm.
Connection active time total Amount of time elapsed while a thread makes a connection active.
Connection charset time total Amount of time elapsed while checking a connection's character set.
Connection fadeout freetime total Amount of time elapsed while this thread's free connections are marked for fadeout. Note that this thread is probably *not* the thread doing the work. Most likely it is the Alarm Handler doing the work.
Connection fadeout time total Amount of time elapsed while this thread's assigned connections are marked for fadeout. Note that this thread is probably *not* the thread doing the work. Most likely it is the Alarm Handler doing the work.
Connection give-up total Number of times the thread gave up a connection.
Connection give-up time total Amount of time elapsed while a thread gives up a connection.
Connection idle time total Amount of time elapsed while a thread makes a connection idle.
Connection open total Number of times the thread made a connection request and received one that had to be opened.
Connection request total Number of times the thread made a connection request and received one that was already open.
Connection sort order time total Amount of time elapsed while checking a connection's sort order.
Connection time total Amount of time elapsed while a thread obtains a connection.
Message queue get time total Number of centiseconds spent getting a message from an internal message queue.
Message queue put time total Number of centiseconds spent putting a message on an internal message queue.
Sleep time total Amount of time elapsed after a thread puts itself to sleep.
Yield time total Amount of time elapsed after a thread yields the CPU.
Blocks read total Number of blocking (SQM_WAIT_C) reads that an RSI thread performed against SQM thread that manages a RSI queue.
Bytes sent total Bytes that an RSI sender thread delivered.
Messages sent total RSI messages that an RSI thread sent. These messages contain the distribute command.
Packets sent total Packets that an RSI sender thread sent.
Send packet time total Time, in milliseconds, spent in sending packets of data to the RRS.
Bytes received total Bytes that an RSI user thread received. This size includes the TDS header size when in 'passthru' mode.
Command length total Length of an RSI command.
Commands received total Commands that an RSI user thread received. Includes RSI message, get truncation requests, etc.
Execution time total The amount of time, in milliseconds, RSI User thread is scheduled by OCS.
Receive packet time total The amount of time, in milliseconds, spent receiving network packets.
GetTrunc send time total The amount of time, in milliseconds, spent responding to 'get truncation' requests.
SetTrunc send time total The amount of time, in milliseconds, spent responding to 'set truncation' requests.
Queue
Queue Writes The number of commands written to the queue in the Replication Server. Monitoring Statistics
Queue Writes per Second The number of commands written to the queue in the Replication Server per second.
Queue Reads The number of commands read from the queue in the Replication Server.
Queue Reads per Second The number of commands read from the queue in the Replication Server per second.
Queue Bytes The number of bytes written to the queue in the Replication Server.
Queue Bytes per Second The number of bytes written to the queue in the Replication Server per second.
Queue Size The size of the queue in the Replication Server.
Queue SQT Cache Percent Used The percent of SQT cache size used for inbound queue.
Duplicates Transaction during monitoring interval The number of duplicate commands detected and ignored by the SQM.
Transactions removed per second The number of transactions removed from the SQT cache per second.
SQT cache low boundry The smallest size to which SQT cache would be configured before transactions start being removed from cache.
Commands processed by DIST per second The number of commands that the distributor thread processed for an inbound queue per second.
Commands processed without rep def per second The number of commands that did not have an associated replication definition, which the distributor thread processed for an inbound queue per second.
Percent of blocks that were full when written to the queue The percentage of blocks that were full when written to the queue.
Percent of reads that came directly from the cache The percentage of reads that came directly from the cache.
Blocks full write total Number of full blocks that an SQM thread wrote. Individual blocks can be written due either to block full state or to sysadmin command slow_queue (only one message per block). Performance Counters
Blocks written total Number of 16K blocks that an SQM thread wrote to a stable queue.
Bytes written total Bytes that an SQM thread wrote to a stable queue.
Commands written total Commands that an SQM thread wrote to a stable queue.
Command size total Command size written to a stable queue.
Segments active total Active segments of an SQM queue: the number of rows in rs_segments for the given queue where used_flag = 1.
Queue segments allocated total Segments allocated to a queue during the current statistical period.
Queue segments deallocated total Segments deallocated from a queue during the current statistical period.
Time new segments allocated total The elapsed time, in milliseconds, to allocate a new segment. Timer starts when a segment is allocated. Timer stops when the next segment is allocated.
Updates to rs_oqid total Updates to the RSSD. rs_oqid table by an SQM thread. Each new segment allocation may result in an update of rs_oqid value stored in rs_oqid for recovery purposes.
Write requests total Message writes that an SQM client requested.
Write timer pop total SQM writer thread initiated a write request due to timer expiration.
Blocks read total Number of 16 blocks that an SQM Reader (SQMR) thread read from a stable queue.
Blocks read cached total Number of 16K blocks that an SQMR thread read from a cache.
Commands read total Commands that an SQMR thread read from a stable queue.
Time waiting for queue write total The amount of time SQMR wait for a queue write.
Unread blocks total The number of blocks within a partially read segment that are yet to be read.
Unread segments total The number of segments yet to be read.
Read time total The amount of time taken for SQMR to read a block.
Memory used total SQT thread memory use. Each command structure allocated by an SQT thread is freed when its transaction context is removed. For this reason, if no transactions are active in SQT, SQT cache usage is zero.
Transactions added closed total Transactions added to the Closed queue.
Transactions removed closed total Transactions removed from the Closed queue.
Commands read total Commands read from SQM. Commands include XREC_BEGIN, XREC_COMMIT, XREC_CHECKPT.
Commands in tran total Commands in transactions completely scanned by an SQT thread.
Transactions added total Transactions added to the Open queue.
Transactions removed total Transactions removed from the Open queue.
Transactions added Read Q total Transactions added to the Read queue.
Transactions removed Read Queue total Transactions removed from the Read queue.
Cache lower bound size total The smallest size to which SQT cache can be configured before transactions start being removed from cache.
Closed transactions total Current closed transaction count.
Open transactions total Current open transaction count.
Read transactions total Current read transaction count.
Truncation queue transaction total Current truncation queue transaction count.
Transactions removed total Transactions whose constituent messages have been removed from memory. Removal of transactions is most commonly caused by a single transaction exceeding the available cache.
Truncation queue transactions added total Transactions added to the Truncation queue.
Truncation queue transactions removed total Transactions removed from the Truncation queue
Partition
Partition State The state of the partition in the Replication Server. Monitoring Statistics
Partition Usage The percent of the partition in the Replication Server that is currently in use.
RepAgent
Replication Agent State The state of the Replication Agent. RA Monitoring Statistics
RepAgent VM
VM maximum memory Maximum memory (in bytes) available to the Java Virtual Machine (VM). Performance Counters
VM total memory allocated Total memory (in bytes) allocated to the Java VM at start-up.
VM free memory Memory (in bytes) allocated but not used by the Java VM.
VM memory usage Memory (in bytes) allocated and in use by the Java VM.
VM % max memory used Percentage of the the maximum memory available to the Java VM, currently in use by the Java VM.
RepAgent LTI Statistics
Number of LTL commands sent Total number of LTL commands sent to Replication Server. Performance Counters
LTL Commands Sent per Second LTL commands sent to Replication Server per second.
Avg LTL command size Average size (in bytes) of each LTL command sent to Replication Server.
Avg LTL commands/sec Average number of LTL commands sent per second to Replication Server.
Total bytes sent Number of bytes sent to Replication Server.
Avg Bytes/second during transmission Average bytes per second sent over connection to Replication Server.
Avg LTL buffer cache time Average time (in milliseconds) it takes between placing the LTL commands into the LTL buffer to the time it is actually sent to the Replication Server.
Avg Rep Server turnaround time Average time (in milliseconds) it takes Replication Server to acknowledge each LTL command buffer sent.
Avg data arrival time Average time (in milliseconds) log transfer interface (LTI) waits between receiving change sets from Log Reader.
Avg time to create distributes Average time (in milliseconds) LTI takes to convert a change-set into LTL.
Avg LTL buffer size Average size (in bytes) of each LTL buffer sent to Replication Server
Avg LTM buffer utilization (%) Average utilization (in percentage of LTL buffer size) of each LTL buffer sent to Replication Server.
Avg LTL commands/buffer Average number of LTL commands per buffer sent to Replication Server.
LTI encoded column name cache size Encoded column name cache size.
LTI input queue size Number of change sets in the LTI input queue
LTI output queue size Current number of LTL commands in the LTI output queue.
RepAgent Log Reader Statistics
Total operations scanned Number of operations read from the database transaction log. Performance Counters
Operations Scanned per Second Number of operations read from the database transaction log per second.
Total operations processed Number of operations processed from the database transaction log.
Operations Processed per Second Number of operations process from the database transaction log per second.
Total operations skipped Number of operations read from the database transaction log and skipped for any reason.
Operations Skipped per Second Number of operations read from the database transaction log and skipped for any reason per second.
Current operation queue size Current size (in bytes) of the Log Reader input queue.
Current scan buffer size Current size (in bytes) of the Log Reader scan buffer.
Total maintenance user operations filtered Number of maintenance-user operations read from the database transaction log and skipped.
Total Maintenance User Operations Filtered per Second Number of maintenance-user operations read from database transaction log and skipped per second.
RepAgent LTM Statistics
Items held in Global LRU Cache Number of object references in the internal Least Recently Used cache. Performance Counters
Adaptive Server RepAgent Thread
ASE Replication Agent Thread Status The status of the Adaptive Server RepAgent thread in the data server. RepAgent Statistics
ASE Replication Agent Thread Log The difference between the current marker and the end of the log.
ASE Replication Agent Thread Operations Scanned The number of operations scanned from the Adaptive Server transaction log since the last monitoring cycle.
ASE Replication Agent Thread Commands Processed The number of commands processed and sent to the Replication Server since the last monitoring cycle.
Key Performance Areas
KPA Description
KPA_RS_PERF_STATS Replication Server common statistics.
KPA_RS_HEATCHART Replication Server heat chart statistics.
kpa_path Replication path statistics.
KPA_PATH_RS_TICKET The latency times at various points in the path between the primary database and the replicate database.
kpa_connection Replication Server connection statistics.
KPA_CONNECTION_PERF_STATS Replication Server connection performance statistics.
kpa_logical_connection Replication Server logical connection statistics.
kpa_route Replication Server route statistics.
KPA_ROUTE_PERF_STATS Replication Server route performance statistics.
kpa_queue Replication Server queue statistics.
KPA_QUEUE_PERF_STATS Replication Server queue performance statistics.
kpa_partition Replication Server partition statistics.
Availability Statistics Replication Agent availability statistics.
VM Statistics Replication Agent performance statistics on JAVA VM component.
LTM component Statistics Replication Agent performance statistics on log transfer manager (LTM) component.
LTI component Statistics Replication Agent performance statistics on LTI component.
LR component Statistics Replication Agent performance statistics on log reader (LR) component.
kpa_ase Adaptive Server statistics
kpa_rat_ctrl_availability Adaptive Server RepAgent thread controller availability statistics.
kpa_rat_availability Adaptive Server RepAgent thread availability statistics.
Related concepts
About Statistics
Alerts
Related tasks
Creating an Alert
Assigning a Role to a Login or a Group
Configuring the E-mail Server