腾讯云可观测平台 云数据库 PostgreSQL 监控指标

这篇列出腾讯云PostgreSQL实例的监控指标,包括活跃连接数、CPU利用率、QPS、存储空间等,并说明各指标的单位、维度和统计规则,用于监控实例性能与健康状态。 – 核心指标:涵盖连接数、CPU、内存、存储、QPS、TPS等关键性能指标。 – 统计规则:每个指标支持多种统计粒度(如5秒、60秒、300秒)和统计类型(如first、max、sum)。 – 入参说明:查询监控数据需指定Namespace为QCE/PO。

相关产品
腾讯云可观测平台 产品页,含版本规格、价格与购买入口。

一、命名空间

Namespace = QCE/POSTGRES

二、监控指标

指标英文名 指标中文名 说明 单位 维度 统计规则
[period, statType]
2pc 2pc 事务数 对数据库发起采集时,当前的2pc 事务数量 Count resourceId [ 5s, first ]
[ 60s, sum ]
[ 300s, sum ]
ActiveConns 活跃连接数 对数据库发起采集时,数据库瞬时活跃连接(非 idle 连接) Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
ClusterInFlow 网络入流量 数据库实例监听端口上的入流量 KBytes/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
ClusterOutFlow 网络出流量 数据库实例监听端口上的出流量 KBytes/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
Connections 连接数 对数据库发起采集时,数据库当前总连接数 count resourceId [ 5s, first ]
[ 60s, avg ]
[ 300s, avg ]
[ 3600s, avg ]
[ 86400s, avg ]
ConnUtilization 连接利用率 当前实例连接数 / 最大连接数 % resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
Cpu CPU 利用率 CPU 实际利用率 % resourceId [ 5s, first ][ 60s, max ][ 300s, max ]
DataFileSize 数据文件大小 数据文件占用空间大小 KBytes resourceId [ 5s, max ]
[ 60s, max ]
[ 300s, max ]
Deadlocks 死锁数 一个统计周期内的所有死锁数 Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
FlushLatency 备库日志落盘延迟 备库落盘 LSN 与主实例当前 LSN 之间的大小差异,对于主实例而言,此指标可以体现出故障切换的数据丢失大小,对于只读实例而言,此指标可以体现出数据延迟大小 Bytes resourceId [ 60s, max ]
[ 300s, max ]
HitPercent 缓冲区缓存命中率 一个统计周期内的所有 SQL 语句执行的命中率 % resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
[ 3600s, max ]
[ 86400s, max ]
IdleConns 空闲连接数 对数据库发起采集时,查询出来的数据库瞬时空闲连接(idle 连接) Count resourceId [ 5s, first ]
[ 60s, avg ]
[ 300s, avg ]
IdleInXact 空闲事务数 对数据库发起采集时,数据库正在处于 idle in transaction 状态的事务数量 Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
LogFileSize 日志文件大小 wal 日志文件占用空间大小 KBytes resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
Long2pc 超过5s未提交的2pc 事务数 对数据库发起采集时,当前执行时间超过5s的2pc 事务数量 Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
LongIdleInXact 超过5秒的空闲事务数 对数据库发起采集时,状态处于idle in transaction且时间超过5秒的事务数量 Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
LongQuery 执行时长超过1秒的 SQL 数 对数据库发起采集时,正在执行且执行时间超过1s的 SQL 数量 Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
LongWaiting 等待超过5s的会话数 对数据库发起采集时,数据库正在等待且超过5秒的会话数量(状态为 waiting) Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
LongXact 执行时长超过1秒的事务数目 对数据库发起采集时,执行时间超过1秒的事务数量 Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
Memory 内存占用量 内存实际使用量 MBytes resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
MemoryRate 内存使用率 内存实际使用率 % resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
NewConnIn5s 5秒内新建连接数 对数据库发起采集时,查询出最近5秒内建立的所有连接数 Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
OtherCalls 其他请求数 一个统计周期内的其他请求数(除 SELECT、INSERT、UPDATE、DELETE 以外的请求数) Count resourceId [ 5s, first ]
[ 60s, sum ]
[ 300s, sum ]
[ 3600s, sum ]
[ 86400s, avg ]
Qps 每秒查询数 平均每秒执行的 SQL 语句数量 count/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
[ 3600s, max ]
[ 86400s, max ]
ReadCalls 读请求数 一个统计周期内的读请求数 Count resourceId [ 5s, first ]
[ 60s, sum ]
[ 300s, sum ]
[ 3600s, sum ]
[ 86400s, max ]
ReadWriteCalls 读写请求数 一个统计周期内的读写请求数 Count resourceId [ 5s, first ]
[ 60s, sum ]
[ 300s, sum ]
[ 3600s, sum ]
[ 86400s, avg ]
RemainXid 剩余 XID 数量 对数据库发起采集时,显示当前剩余 xid 最少的库的剩余 xid 数量。只读实例无此指标 count resourceId [ 5s, first ]
[ 60s, min ]
[ 300s, min ]
[ 3600s, min ]
[ 86400s, min ]
ReplayLag 主备数据同步延迟时间 日志从主库发送至备库与备库接收到日志并回放之间的时间差异。实例版本在10.x 以上版本才有此指标 s resourceId [ 60s, max ]
[ 300s, max ]
SlaveApplyDelay 主备数据同步延迟 备库回放 LSN 与主实例当前 LSN 之间的大小差异,对于主实例而言,此指标可以体现出故障切换的 RTO,对于只读实例而言,此指标可以体现出数据延迟大小 Bytes resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
SlowQueryCnt 慢查询数量 一个统计周期内的慢查询个数 Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
SqlRuntimeAvg 平均执行时延 一次统计周期内所有 SQL 语句的平均执行时延 ms resourceId [ 5s, first ]
[ 60s, avg ]
[ 300s, avg ]
[ 3600s, avg ]
[ 86400s, avg ]
SqlRuntimeMax 最长 TOP10执行时延 一次统计周期内最长 TOP10的 SQL 平均执行时延 ms resourceId [ 5s, first ]
[ 60s, avg ]
[ 300s, avg ]
[ 3600s, avg ]
[ 86400s, avg ]
SqlRuntimeMin 最短 TOP10执行时延 一次统计周期内最短 TOP10的 SQL 平均执行时延 ms resourceId [ 5s, first ]
[ 60s, avg ]
[ 300s, avg ]
[ 3600s, avg ]
[ 86400s, avg ]
Storage 已用存储空间 整实例空间占用大小 GB resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
[ 3600s, max ]
[ 86400s, max ]
StorageRate 存储空间使用率 总的存储空间使用率,包括临时文件、数据文件、日志文件以及其他类型的数据库文件 % resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
[ 3600s, max ]
[ 86400s, max ]
TempFileSize 临时文件大小 临时文件的大小 KBytes resourceId [ 5s, max ]
[ 60s, max ]
[ 300s, max ]
Throughput 吞吐率 实例每秒读写磁盘速度 KBytes/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
ThroughputRead 读吞吐率 实例每秒读磁盘速度 KBytes/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
ThroughputWrite 写吞吐率 实例每秒写磁盘速度 KBytes/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
Tps 每秒事务数 平均每秒执行成功的事务数(包括回滚和提交) Count/s resourceId [ 5s, first ]
[ 60s, avg ]
[ 300s, avg ]
TupDeleted 每秒删除记录数 一个统计周期内的平均每秒删除的 tuple 数量 Count/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
TupFetched 每秒索引扫描记录数 一个统计周期内的平均每秒索引扫描的 tuple 数量 Count/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
TupInserted 每秒插入记录数 一个统计周期内的平均每秒插入的 tuple 数量 Count/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
TupReturned 每秒全表扫描记录数 一个统计周期内的平均每秒全表扫描的 tuple 数量 Count/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
TupUpdated 每秒更新记录数 一个统计周期内的平均每秒更新的 tupe 数量 Count/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
Waiting 等待会话数 对数据库发起采集时,数据库正在等待的会话数量(状态为 waiting) Count resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
WriteCalls 写请求数 一个统计周期内的写请求数 Count resourceId [ 5s, first ]
[ 60s, sum ]
[ 300s, sum ]
[ 3600s, avg ]
[ 86400s, avg ]
XactCommit 事务提交数 一个统计周期内的平均每秒提交事务数 Count/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
XactRollback 事务回滚数 一个统计周期内的平均每秒回滚事务数 Count/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
XlogDiff 备库日志发送与回放位置差异 日志从主库发送至备库与备库回放完成之间的大小差异,主要反映了备库日志应用的速度,主要能够通过此指标查看出备库的性能、网络传输的速度。只读实例无此指标 Byte/s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]
[ 3600s, max ]
[ 86400s, max ]
XlogDiffTime 备库日志落盘时间延迟 日志从主库发送至备库与备库接收到日志并落盘之间的时间差异。只读实例无此指标,且实例版本在10.x 以上版本才有此指标 s resourceId [ 5s, first ]
[ 60s, max ]
[ 300s, max ]

注意:每个指标的统计粒度(Period)可取值不一定相同,可通过 DescribeBaseMetrics 接口获取每个指标支持的统计粒度。

三、各维度对应参数总览

参数名称 维度名称 维度解释 格式
Instances.N.Dimensions.0.Name resourceId resourceId 维度名称 输入 String 类型维度名称:resourceId
Instances.N.Dimensions.0.Value resourceId 实例具体的 resourceId 输入实例的具体 resourceId,例如:postgres-123456

四、入参说明

查询 PostgreSQL 监控数据,入参取值如下:
&Namespace=QCE/POSTGRES
&Instances.N.Dimensions.0.Name=resourceId
&Instances.N.Dimensions.0.Value=实例的 resourceId

赞 (0)
上一篇 4天前
下一篇 4天前