When influxdb counts cardinality in a Measurement, 
it counts the combination of all tags. 

For example, if my measurement has the following tags: 
3 os, 200 devices, 3 browsers, 
then the cardinality is > 3x200x3=1800. 

-- show estimated cardinality of the series on current database
SHOW SERIES CARDINALITY

-- show estimated cardinality of the series on specified database
SHOW SERIES CARDINALITY ON mydb

-- show exact series cardinality
SHOW SERIES EXACT CARDINALITY

-- show series cardinality of the series on specified database
SHOW SERIES EXACT CARDINALITY ON mydb

> SHOW SERIES CARDINALITY on appMetrics
cardinality estimation
----------------------
105336
> SHOW SERIES CARDINALITY on hardWareMetrics
cardinality estimation
----------------------
10703

> SHOW SERIES EXACT CARDINALITY ON hardWareMetrics
name: burrow_group
count
-----
112

name: disk
count
-----
14256

name: diskio
count
-----
3920

name: influxdb_database
count
-----
40

name: influxdb_httpd
count
-----
8

name: influxdb_shard
count
-----
1905

select sum(pc) from advertise
where publisher=‘atlasdata’ and platform = ‘baidu’
and time>='2018-08-01' and time <'2018-09-01'

InfluxDB stores measurement and tag information in an index so data can be queried quickly.
In earlier versions, the index was stored in-memory, requiring a lot of RAM and restricting the number of series that a machine could hold (typically, 1-4 million series, depending on the machine).
Time Series Index (TSI) stores index data both in memory and on disk, removing RAM restrictions. This lets you store more series on a machine. TSI uses the operating system’s page cache to pull hot data into memory, leaving cold data on disk.

index-version = "inmem"

inmem:基于内存

tsi1：基于磁盘


The default (inmem) index is an in-memory index that is recreated at startup. To enable the Time Series Index (TSI) disk-based index, set the value to tsi1.

> show stats
name: runtime
Alloc      Frees     HeapAlloc  HeapIdle   HeapInUse  HeapObjects HeapReleased HeapSys    Lookups Mallocs   NumGC NumGoroutine PauseTotalNs Sys        TotalAlloc
-----      -----     ---------  --------   ---------  ----------- ------------ -------    ------- -------   ----- ------------ ------------ ---        ----------
1453852384 218795736 1453852384 1682595840 1571332096 9628928     0            3253927936 178461  228424664 96    334          784143583    3418348024 55161719256

name: queryExecutor
queriesActive queriesExecuted queriesFinished queryDurationNs recoveredPanics
------------- --------------- --------------- --------------- ---------------
1             73169           73168           27803545203003  0


name: database
tags: database=_internal
numMeasurements numSeries
--------------- ---------
12              1283

name: database
tags: database=metrics
numMeasurements numSeries
--------------- ---------
862             105336

name: database
tags: database=telegraf
numMeasurements numSeries
--------------- ---------
40              10703

# 详细配置说明见官方文档
# https://docs.influxdata.com/influxdb/v1.8/administration/config/#data-settings
 
[data]
  #说明: wal预写日志log,用于事务一致性
  #默认为0，每次写入都落盘。
  #修改为1s, 根据业务场景，不保证强一致性,可采用异步刷盘
  #[优化点]:用于减轻磁盘io压力
  wal-fsync-delay = "1s"
   
  #说明: influx索引
  #默认为inmem,创建内存型索引,在delete retention会消耗过高内存
  #修改为tsi1, 注意重建tsi1索引(https://blog.csdn.net/wzy_168/article/details/107043840)
  #[优化点]:降低删除保留策略时的内存消耗
  index-version = "tsi1"
   
  #说明: 压缩TSM数据,一次落盘的吞吐量
  #默认48m
  #修改为64m
  #[优化点]:增大写入量，减轻io压力
  compact-throughput = "64m"