지금 지원 담당자와 채팅
지원 담당자와 채팅

Foglight for SQL Server (Cartridge) 5.9.4.10 - User Guide

Introduction to this Guide Using Foglight for SQL Server
Viewing the Databases Dashboard Overview Dashboard Advisories Dashboard Monitoring Data Replication Monitoring SQL Performance Reviewing Memory Usage Reviewing the Instance Activity Reviewing Database Usage Reviewing the Services Using the HADR Drilldown Using the Logs Drilldown Reviewing Configuration Settings Viewing User-defined Performance Counters and Collections
Monitoring Business Intelligence Services Administering Foglight for SQL Server
Configuration Settings Managing Foglight for SQL Server Agent Settings Reviewing Foglight for SQL Server Alarms Generating Reports Monitoring SQL Server instances on VMware servers
Glossary
A
Access methods Adhoc SQL Plans Alarm Alert Allow updates Anonymous subscription Authentication AutoClose AutoGrow Automatic Discovery AutoShrink B Batch BCP (Bulk Copy Program) Blocking Books Online Bound trees Buffer Buffer cache Buffer pool Bulk copy Bulkinsert Bulk load C Cache CAL Calibration Cardinality Cartridge Chart Checkpoint process Client network utility CLR Compile Connect Connection Connectivity software CPU Usage Cursors D Data access components Data file Data storage engine Database Database object DataFlow DBCC DBID DBO Deadlock Disk queue length Disk transfer time Disk utilization DiskPerf Distributing instance Distributor DMO Drilldown DTC DTS E Error log Event alert Execution contexts Extended stored procedures Extent External procedures F File Filegroup File cache Foglight Agent Manager Foglight Management Server Foreign key Forwarded records Free buffers Free list Free pages Free potential Free space Freespace scans Full text search G GAM Ghosted records Growth increment GUID H Hash buckets Hashing Heap Hit rate Hit ratio Host name Host process I I/O Index Indid Integrated security Intent Locks J Job K Kernel memory Kernel mode Kill L Latch Lazy writer Licensing Lightweight pooling Lock Lock area Lock escalation Lock mode Log Log cache Log writer Logical I/O LRU LSN M Master MaxSize MDAC Metric Misc.normalized trees Model Monitor page file N Named pipes Net library NIC Null O OBID Object plans OLAP OLAP service OLTP Optimizer Optimizer cache osql P Page life expectancy Paging Panel Parse Parser Per seat licensing Per server licensing Performance alert Physical I/O Physical read Physical write PID Pinned Plan Plan cache Potential growth Prepared SQL plans Primary key Privileged mode Procedure cache Procedure plans Process Profiler Publication database Publisher Publisher databases Publishing server Pull subscription Pulse Push subscription Q Query plan R RAID Random I/O Read ahead Recompile Referential integrity Relational data engine Replication procedure plans Role Rollback S sa Schema locks Sequential I/O Session Severity SGAM Shared locks Show advanced options SMP Sort, Hash, Index Area SPID Spike Spinner SQL Agent Mail SQL Mail SQL Plans SQL Server Agent SQL Server authentication SQL Server books online Standard deviation Stolen pages Stored procedure Support service SYSADMIN role T TDS TempDB Temporary tables and table variables Threshold Torn page detection Transaction Trigger Trigger plans Truncate Trusted U UMS Unused space User connection area User mode V Virtual log file VLF W Waitfor Windows authentication mode Working set
Reference
SQL PI Repository Cold Backup Procedure SQL Performance Investigator Metrics
Active Time All SQL Agents CPU Usage All SQL Agents Resident Memory Usage Availability Average Physical I/O Operations Average SQL Response Time Backup Recovery Wait Blocked Lock Requests Checkpoint Pages CLR Wait CPU Usage CPU Wait Cursor Synchronization Wait Database Replication Wait Deferred Task Worker Wait Degree of Parallelism Disk Utilization DTC CPU Usage DTC Resident Memory Usage Distributed Transaction Wait Executions Ended Executions Started External Procedures Wait Full Scans Full Text Search CPU Usage Full Text Search Resident Memory Usage Full Text Search Wait Free Buffer Wait Hosted Components Wait IO Bulk Load Wait IO Completion Wait IO Data Page Wait IO Wait Latch Buffer Wait Latch Wait Latch Savepoint Wait Lazy Writes Lock Wait Lock Bulk Update Wait Lock Exclusive Wait Lock Intent Wait Lock Requests Lock Schema Wait Lock Shared Wait Lock Update Wait Lock Wait Log Buffer Wait Log Flushes Log Other Wait Log Synchronization Wait Log Wait Log Write Wait Memory Wait Network IO Wait Network IPC Wait Network Mirror Wait Network Wait Non SQL Server CPU Usage Non SQL Resident Memory Usage OLAP CPU Usage OLAP Resident Memory Usage OLEDB Provider Full Text Wait Other CPU Usage Other Miscellaneous Wait Other Wait Overall CPU Page Life Expectancy Page Splits Parallel Coordination Wait Physical I/O Physical Memory Used Physical Page Reads Physical Page Writes Probe Scans Plan Cache Hit Rate Range Scans Rec Ended Duration Remote Provider Wait Run Queue Length Samples Service Broker Wait Session Logons Session Logoffs SQL Agent CPU Usage SQL Agent Resident Memory Usage SQL Executions SQL Mail CPU Usage SQL Mail Resident Memory Usage SQL Recompilations SQL Response Time SQL Server Background CPU Usage SQL Server Cache Memory SQL Server Connections Memory SQL Server Connections Summary SQL Server Foreground CPU Usage SQL Server Resident Memory Usage SQL Server Swap Memory Usage Synchronous Task Wait Table Lock Escalation Target Instance Memory Total CPU Usage Total Instance Memory Virtual Memory Used
Rules Collections and Metrics

Investigating Resource Pool Memory Issues

The Resource Pools (RP) metrics are supported only beginning with SQL Server v2008.

The middle division of the Memory Summary dashboard provides metrics which may help you to prevent resource pool memory related errors. For example:

The first chart in the Resource Pool section displays the resource pools available for application and server usage and alerts memory utilization if occurs on any resource pool. The values are grayed out when the resource governor is disabled.

The list contains the following values:

Two resource pools "internal" and "default" are created when SQL Server is installed. Clicking on the resource pool displays how the usages changes over time.

Further drilling into the databases related to this resource pool is also available as of SQL Server v2014, for databases which have the In-Memory feature enabled. When using In-Memory databases, it is highly recommended to define a dedicated resource pool.

The lower part of the Resource Pool section provides metrics on up to eight resource pools over time. Click the Properties tab to view the definition parameters of the resource pools on the server

Investigating Cached Memory Issues

The Memory Plan metrics provide data on cache-related issues.

Plan cache is a component of SQL memory that stores query plans for re-use. When a query is issued against SQL, the optimizer attempts to re-use a cached plan if the traits of the query permit - but can only do so if that plan resides in cache, otherwise it needs to compile a new plan.

Low hit ratio may indicate a lack of memory resources, a new plan might also be compiled even if a similar plan exists in cache when:

Plan compilations are expensive though; if plans are flushed from cache only to be recompiled afterwards because of size constraints even more memory resources are consumed.

Selecting the Plan Cache hit ratio allows you to investigate changes over time. Further drilling down leads to the Plan Cache dashboard.

Buffer cache (Also called Buffer Pool) is the place in system memory that holds data and index pages read from disk. This has two purposes:

Buffer Cache size is determined among other things by server memory and the target server memory specified in the Max Server Memory parameter. When that threshold is reached and SQL Server needs to read more pages, previously cached pages are discarded.

A low hit ratio may indicate lack of memory. It also indicates that the Buffer Cache cannot improve performance by reducing IO operations. As such, the Buffer Cache size can correlate to improved performance during a heavy workload.

Improving Buffer Cache hit ration in SQL Server 2014 is also possible by setting a Buffer Pool Extension on an SSD drive.

The hit rate for this cache should normally be above 90%. Selecting the Buffer cache hit ratio bar allows you to investigate these metrics over time. Further drilling down leads to the Buffer Cache dashboard.

New in SQL Server 2014, In-Memory OLTP can significantly improve OLTP database application performance.

The Memory Usage of In-Memory (XTP) indicates the amount of data used by memory optimized tables. A large percentage indicates a server that heavily uses In-Memory OLTP features. It is recommended that you dedicate a specific resource pool to databases containing In-Memory OLTP objects hence ensuring proper allocation of memory resources among applications on the SQL Server.

The Oldest XTP Transaction currently active on the server indicates issues regarding long running XTP transactions. Any value over 20 seconds is regarded a warning to an un-healthy behavior on the server. Long-running transactions should be avoided with memory-optimized tables. Such transactions increase the likelihood of conflicts and subsequent transaction terminations. A long-running transaction also defers garbage collection.

The longer a transaction runs, the longer In-Memory OLTP keeps recently deleted row versions, which can decrease lookup performance for new transactions.

Selecting the Oldest XTP transaction displays general information regarding the transaction and the session that had initiated it.

The Major Memory Areas chart displays the amount of memory allocated to the Buffer cache, Plan cache, free memory and XTP. Use this chart to ensure that:

The Minor Memory Areas chart displays the main areas SQL Server uses for carrying out database operations within memory. The display is broken down into each of the main cache areas:

Use this chart to review the efficiency of the memory management, as reflected in the amount of memory allocated to each area.

Monitoring Buffer Cache-related Data

The Buffer Cache panel allows investigating the utilization of the instance’s buffer cache, by reviewing this cache’s top-consuming objects, the page allocation, and the buffer cache hit rate.

Reviewing the Top-consuming Objects of the Buffer Cache

The Buffer Cache table displays the objects that currently occupy the most space in the buffer cache (the Top N buffer cache objects).

To improve performance, Foglight for SQL Server limits the number of records that can be displayed in this table to 300 rows, but by default, a maximum of 20 records are displayed.

The name of the database where the object resides.

The user name of the object's owner.

The table name (if the object is a table).

The index name (if the object is an index).

If this column is left empty, the object is a heap.

The amount of cache used by the object.

The index identifier (if the object is an index).

This metric can have one of the following values:

The file group where the object resides.

The percentage of object in the cache.

The percentage of buffer cache used by the object.

Because the process of determining the largest objects in the cache is highly CPU-intensive, this table can take some time to populate.

관련 문서

The document was helpful.

평가 결과 선택

I easily found the information I needed.

평가 결과 선택