Monitoreo de Bases de Datos
Monitoreo de Bases de Datos Las bases de datos son pieza fundamental de una Infraestructura, es de vital importancia su correcto monitoreo de métricas para efectos de lograr su mejor desempeño y disponibilidad. El monitoreo se realiza usando dos puntos de vista: 1. Servidor (monitoreo de integridad del servidor de base de datos) 2. Usuario final (monitoreo de tiempos de respuesta para querys SQL específicos) La meta final es asegurar la disponibilidad y un desempeño óptimo de las bases de datos de manera de no comprometer la productividad del negocio y satisfacción del usuario final. Funcionalidades: Monitoreo en tiempo real 7x24 Monitoreo de disponibilidad, desempeño y cumplimiento de SLA Reportes históricos de desempeño Soporte a bases de datos: o Oracle o MS SQL o DB2 o Informix o MySQL o Sybase
Métricas Monitoreo Cada base de datos tiene sus métricas específicas, a continuación y listado de dichas métricas: Oracle Database uptime. Tablespace growth. Database growth. Tablespace status. Index status. Data file status. Rollback segment status. Fragmented segments. The extents that can't extend. The data dictionary cache hit ratio. The data buffer cache hit ratio. The Redo copy latch hit ratio. The library cache hit ratio. The sort hit ratio. PGA resource consumption (monitor memory consumption of Oracle users). The rollback segment contention. The number of invalid objects. The number of chained rows. The number of users currently logged onto the server. The MTS response time. The number of MTS waits. The enqueue resources. The UGA memory usage. User locks and locked users. Lock waits event time. The user buffer cache hit ratio. System waits and user waits. Datafile i/o. System statistics. Global cache service utilization for RAC. Global cache fusion ratio for RAC. Global cache lock get time for RAC. Global cache lock conversion timeouts for RAC. Global cache average lock get time for RAC. Global cache corrupt blocks count for RAC. Global cache lost blocks count for RAC. Long running queries.
Microsoft SQL Server Database uptime. Database state / status. Number of databases available. The data file size for each database. Log file size for each database. File group size for each database. Table size for each database. The buffer cache hit ratio. The log file cache hit ratio. The number of active users. The number of users currently logged onto the server. Number of deadlocks pr. second. Number of transactions pr. second. Number of database page reads/writes pr. second. Number of flush waits pr. second. Number of latch requests pr. second. Number of full scans (table or index) pr. seconds. The usage (growth/shrinking) of the transaction logs Table/index fragmentation. Memory resources. CPU and I/O resources. Locking and locking resources. Free connections. Backup status. Long running queries (SQL Server 2005 only). Long running jobs.
MySQL GENERAL Available Databases Bytes Received Bytes Sent Max Per Thread Memory Max Shared Memory Configured Max Total Memory CONNECTION HISTORY Connection Attempts Successful Connections Refused Connections Connection Refusal Rate Aborted Connections CONNECTIONS Connections Connection Usage Rate Running Threads Peak Open Connections Peak Usage Rate MYISAM KEY CACHE Cache Hit Rate Cache Usage Rate Read Requests Reads From Disk Write Requests Writes To Disk Disk: Total Write Ratio INNODB CACHE Cache Usage Rate Cache Hit Rate Cache Write Waits Rate Latched Pages TABLE LOCKS Locks Acquired Immediately Lock Waits QUERY CACHE Cache Usage Rate Queries In Cache Queries Not Cached Queries Inserted To Cache Cache Hits Queries Pruned From Cache Pruned as percent of Inserts Cache Hit Rate TEMPORARY TABLES Temp Tables Created Temp Tables Created On Disk Disk: Total Ratio BINARY LOG Transactions Used Cache Transactions Saved In Temp File Transactions Saved In Temp File Percent SLOW LAUNCH THREADS Slow Launch Threads SLOW QUERIES Enabled Min Long Query Time Slow Queries SORT BUFFERS Sort Range Sort Scan Temp Files FULL TABLE SCANS Full Table Scan Rate Selects Requiring Full Table Scans Full Table Scans For Joins STATEMENTS All Statements Select Percent
Lock Contention Rate INNODB ROW LOCKS Lock Waits Average Wait Time INNODB LOGS Log Waits Rate Total Size : Cache Size Percent THREAD CACHE Threads Threads Created For Connections Cache Hit Rate TABLE CACHE Tables Open Tables In Cache Rate Table Cache Misses Insert Percent Update Percent Delete Percent Replace Percent Commits Rollback: Commit ratio REPLICATION Slave Running Slave I/O Running Slave SQL running Seconds behind master Last Error Number Slave Retried Transaction Count Open temp tables Master Log Read Position Relay Master Log File Position Relay Log File Position SIZE Table Size
Sybase Server uptime. Local and/or remote monitoring of Sybase Servers. Amount of free log file space. Data file size for each database. Data and log file cache hit ratio. Reports the number of: active users. databases available. users currently logged onto the servers. granted and waited page and table locks. transactions and deadlocks pr. second. database page reads/writes pr. second. disk IO, disk reads and disk writes. committed transactions.
DB2 i_agents_created_ratio i_piped_sorts_rejected db_pool_hit_ratio db_avg_sort_time db_pct_sort_overflows db_avg_sort_heap db_pct_hjs_overflows db_pool_sync_reads db_pool_sync_writes db_pool_sync_idx_writes db_pool_sync_idx_reads db_pool_avg_async_read_time db_pool_avg_async_write_time db_pool_sync_write_time db_pool_avg_write_time db_avg_direct_read_time db_avg_direct_write_time db_cat_cache_hit_rto app_avg_sort_time app_pct_sort_overflows app_pool_hit_ratio app_avg_direct_read_time app_avg_direct_write_time app_cat_cache_hit_rto app_pkg_cache_hit_rto app_locklist_util bp_pool_hit_ratio bp_pool_avg_async_read_time bp_pool_avg_async_write_time bp_pool_sync_write_time bp_pool_avg_write_time bp_avg_direct_read_time bp_avg_direct_write_time bp_pool_sync_reads bp_pool_sync_writes bp_pool_sync_idx_writes bp_pool_sync_idx_reads ts_usable_pages_pct ts_used_pages_pct ts_free_pages_pct ts_max_used_pages_pct
Informix Logging status of the databases Databases for offline status If database is recovering If database is inconsistent If database mirror is offline or If database mirror is recovering Database free space (percent and MB) Buffer cache hit ratio for reads and writes Table extents and table size (MB) Users waiting for locks Chunk I/O read and write statistics Virtual CPU user and system time Connectivity to the database server Uptime (in days) of the database server Reports the number of: databases active users attempts to exceed the maximum number of locks and threads failed attempts to allocate entries in the transaction table times threads had to wait for a latch or a buffer times sessions had to wait for a lock times threads had to wait for critical section checkpoints times potential deadlocks are detected and prevented times distributed deadlock timeout is exceeded while user is waiting for lock foreground writes, LRU writes and chunk writes users waiting for locks Más información: www.agrega.hn Email: info@agrega.hn Tel +50422690133