当前位置：网站首页>Influxdb optimization configuration item

Influxdb optimization configuration item

2022-06-21 20:27:00 【51CTO】

Time series database InfluxDB

InfluxDB Is an open source database for storing and analyzing time series data . It is also the most used time series database .

Default configuration for common options (influxdb.conf)

       
       # 
       
        Disable reporting switch , Default  
       
        Every time  
       
       24 
       
        Hours go by  
       
       usage
       
       .
       
       influxdata
       
       .
       
       com 
       
        Send report 
       
       reporting
       
       -
       
       disabled 
       
       = 
       
       false
       
       # 
       
        For backup and recovery  
       
       RPC 
       
        Binding address of the service 
       
       bind
       
       -
       
       address 
       
       = 
       
       "127.0.0.1:8088"
       
[
       
       meta]
       
       # 
       
        Metadata storage directory 
       
       dir 
       
       = 
       
       "/var/lib/influxdb/meta"
       
       # 
       
       retention 
       
       policy 
       
        Auto create switch 
       
       retention
       
       -
       
       autocreate 
       
       = 
       
       true
       
       # 
       
        Metadata service log switch 
       
       logging
       
       -
       
       enabled 
       
       = 
       
       true
       
[
       
       data]
       
       # 
       
        Data directory 
       
       dir 
       
       = 
       
       "/var/lib/influxdb/data"
       
       # 
       
        Pre - write log directory （write 
       
       ahead 
       
       log）
       
       wal
       
       -
       
       dir 
       
       = 
       
       "/var/lib/influxdb/wal"
       
       # 
       
       fsync 
       
        Write wait time before call , The default is  
       
       0
       
       s,SSD 
       
        Set to  
       
       0
       
       s, Not  
       
       SSD 
       
        Recommended setting as  
       
       0
       
       ms
       
       -
       
       100
       
       ms
       
       wal
       
       -
       
       fsync
       
       -
       
       delay 
       
       = 
       
       "0s"
       
       # 
       
       inmem 
       
        Memory index （ Need a lot of memory ）,tsi1 
       
        Disk based timing index 
       
       index
       
       -
       
       version 
       
       = 
       
       "inmem"
       
       # 
       
        Query log switch 
       
       query
       
       -
       
       log
       
       -
       
       enabled 
       
       = 
       
       true
       
       # 
       
        Fragment cache in reject  
       
       /
       
       write 
       
        The maximum memory size that can be written before the request , The default is  
       
       1
       
       g
       
       cache
       
       -
       
       max
       
       -
       
       memory
       
       -
       
       size 
       
       = 
       
       "1g"
       
       # 
       
       tsm 
       
        The engine caches the snapshot and writes it to  
       
       tsm 
       
        Cache size of the file , Free memory when finished 
       
       cache
       
       -
       
       snapshot
       
       -
       
       memory
       
       -
       
       size 
       
       = 
       
       "25m"
       
       # 
       
        After a while , If the partition does not receive write or delete ,tsm 
       
        The engine will take a snapshot of the cache and write it to the new  
       
       TSM 
       
        file 
       
       cache
       
       -
       
       snapshot
       
       -
       
       write
       
       -
       
       cold
       
       -
       
       duration 
       
       = 
       
       "10m"
       
       # 
       
        After a while , If  
       
       tsm 
       
        The engine did not receive a write or delete ,tsm 
       
        The engine will compress all in the shard  
       
       TSM 
       
        file 
       
       compact
       
       -
       
       full
       
       -
       
       write
       
       -
       
       cold
       
       -
       
       duration 
       
       = 
       
       "4h"
       
       # 
       
       tsm 
       
        Compress the maximum amount of data written to the disk per second , The short burst period can be set to  
       
       compact
       
       -
       
       throughput
       
       -
       
       burst
       
       compact
       
       -
       
       throughput 
       
       = 
       
       "48m"
       
       # 
       
       tsm 
       
        Compress the maximum amount of data written to disk per second during a short burst 
       
       compact
       
       -
       
       throughput
       
       -
       
       burst 
       
       = 
       
       "48m"
       
       # 
       
        Maximum concurrent compression , The default value is  
       
       0 
       
        Will  
       
       50
       
       % 
       
        Of  
       
       CPU 
       
        The number of cores is used to compress 
       
       max
       
       -
       
       concurrent
       
       -
       
       compactions 
       
       = 
       
       0
       
       # 
       
       inmem 
       
        Set item ： Allowed for each database before deleting writes  
       
       series 
       
        Maximum number of , Set up  
       
       0 
       
        There is no limit to 
       
       max
       
       -
       
       series
       
       -
       
       per
       
       -
       
       database 
       
       = 
       
       1000000
       
       # 
       
       inmem 
       
        Set item ： Every  
       
       tag 
       
        Key allowed  
       
       tag 
       
        Maximum number of values , Set up  
       
       0 
       
        There is no limit to 
       
       max
       
       -
       
       values
       
       -
       
       per
       
       -
       
       tag 
       
       = 
       
       100000
       
       # 
       
       tsi1 
       
        Set item ： Prewrite log  
       
       wal 
       
        Threshold size when the file will be compressed into an indexed file 
       
       # 
       
        Smaller thresholds will result in faster compression of log files , And result in lower heap memory usage , But at the cost of write throughput 
       
       # 
       
        Larger thresholds will not compress frequently , Store more... In memory  
       
       series, And provide higher write throughput 
       
       max
       
       -
       
       index
       
       -
       
       log
       
       -
       
       file
       
       -
       
       size 
       
       = 
       
       "1m"
       
       # 
       
       tsi1 
       
        Set item ：tsi 
       
        The index is used to store the previously calculated  
       
       series 
       
        The internal cache size of the result 
       
       # 
       
        Cache results will be quickly returned from the cache , Without matching  
       
       tag 
       
        Subsequent queries of key values are recalculated 
       
       # 
       
        Set this value to  
       
       0 
       
        Caching will be disabled , This can cause query performance problems 
       
       series
       
       -
       
       id
       
       -
       
       set
       
       -
       
       cache
       
       -
       
       size 
       
       = 
       
       100
       
[
       
       coordinator]
       
       # 
       
        Write request timeout 
       
       write
       
       -
       
       timeout 
       
       = 
       
       "10s"
       
       # 
       
        Maximum number of concurrent queries , The default value is  
       
       0 
       
        Means unrestricted 
       
       max
       
       -
       
       concurrent
       
       -
       
       queries 
       
       = 
       
       0
       
       # 
       
        The maximum duration that a query is allowed to execute before it is terminated , The default value is  
       
       0 
       
        Means unrestricted 
       
       query
       
       -
       
       timeout 
       
       = 
       
       "0s"
       
       # 
       
        Maximum duration of slow queries , A query beyond this time will print  
       
       Detected 
       
       slow 
       
       query 
       
        journal 
       
       # 
       
        The default value is  
       
       0
       
       s 
       
        Indicates that slow queries are not recorded 
       
       log
       
       -
       
       queries
       
       -
       
       after 
       
       = 
       
       "0s"
       
[
       
       retention]
       
       # 
       
        Enforce retention policies and retire old data switches 
       
       enabled 
       
       = 
       
       true
       
       # 
       
        Check interval for executing retention policy 
       
       check
       
       -
       
       interval 
       
       = 
       
       "30m0s"
       
[
       
       monitor]
       
       # 
       
        Internal record statistics switch , If there is a large amount of data in the production environment , Proposed closure 
       
       store
       
       -
       
       enabled 
       
       = 
       
       true
       
       # 
       
        A database that records statistics 
       
       store
       
       -
       
       database 
       
       = 
       
       "_internal"
       
       # 
       
        Time interval for recording statistics 
       
       store
       
       -
       
       interval 
       
       = 
       
       "10s"
       
[
       
       http]
       
       # 
       
       http 
       
        switch 
       
       enabled 
       
       = 
       
       true
       
       # 
       
       http 
       
        Binding address 
       
       bind
       
       -
       
       address 
       
       = 
       
       ":8086"
       
       # 
       
        Certification switch 
       
       auth
       
       -
       
       enabled 
       
       = 
       
       false
       
       # 
       
       http 
       
        Request log 
       
       log
       
       -
       
       enabled 
       
       = 
       
       true
       
       # 
       
        Detailed write log switch 
       
       write
       
       -
       
       tracing 
       
       = 
       
       false
       
       # 
       
       flux 
       
        Query switch 
       
       flux
       
       -
       
       enabled 
       
       = 
       
       false
       
       # 
       
       flux 
       
        Query log switch 
       
       flux
       
       -
       
       log
       
       -
       
       enabled 
       
       = 
       
       false
       
       # 
       
       pprof 
       
       http 
       
        switch , For troubleshooting and monitoring 
       
       pprof
       
       -
       
       enabled 
       
       = 
       
       true
       
       # 
       
        Enable  
       
       /
       
       pprof 
       
        Endpoint and bound to  
       
       localhost:
       
       6060
       
       , Used to debug startup performance problems 
       
       debug
       
       -
       
       pprof
       
       -
       
       enabled 
       
       = 
       
       false
       
       # 
       
       https 
       
        switch 
       
       https
       
       -
       
       enabled 
       
       = 
       
       false
       
       # 
       
       https 
       
        certificate 
       
       https
       
       -
       
       certificate 
       
       = 
       
       "/etc/ssl/influxdb.pem"
       
       # 
       
       https 
       
        Private key 
       
       https
       
       -
       
       private
       
       -
       
       key 
       
       = 
       
       ""
       
       # 
       
        The maximum number of rows returned by the query , The default value is  
       
       0 
       
        Allow unlimited 
       
       max
       
       -
       
       row
       
       -
       
       limit 
       
       = 
       
       0
       
       # 
       
        maximum connection , New connections that exceed the limit will be discarded , The default value is  
       
       0 
       
        Disable restrictions 
       
       max
       
       -
       
       connection
       
       -
       
       limit 
       
       = 
       
       0
       
       # 
       
        Client request  
       
       body 
       
        Maximum size of （ In bytes ）, Set to  
       
       0 
       
        Then disable restrictions 
       
       max
       
       -
       
       body
       
       -
       
       size 
       
       = 
       
       25000000
       
       # 
       
        Access log path , If set and when  
       
       log
       
       -
       
       enabled 
       
        When enabled , The request log will be written to this file 
       
       # 
       
        By default, write  
       
       stderr, And  
       
       influxdb 
       
        Logs are mixed together 
       
       access
       
       -
       
       log
       
       -
       
       path 
       
       = 
       
       ""
       
       # 
       
        Request record status filtering , for example  [
       
       "4xx", 
       
       "5xx"]
       
       , Default  [] 
       
        Indicates no filter , All requests are logged and printed 
       
       access
       
       -
       
       log
       
       -
       
       status
       
       -
       
       filters 
       
       = []
       
       # 
       
        Maximum number of concurrent writes , Set to  
       
       0 
       
        Then disable restrictions 
       
       max
       
       -
       
       concurrent
       
       -
       
       write
       
       -
       
       limit 
       
       = 
       
       0
       
       # 
       
        Maximum number of writes queued for processing , Set to  
       
       0 
       
        Then disable restrictions 
       
       max
       
       -
       
       enqueued
       
       -
       
       write
       
       -
       
       limit 
       
       = 
       
       0
       
       # 
       
        The maximum duration of a write waiting in the pending queue 
       
       # 
       
        Set to  
       
       0 
       
        Or will  
       
       max
       
       -
       
       concurrent
       
       -
       
       write
       
       -
       
       limit 
       
        Set to  
       
       0 
       
        Then disable restrictions 
       
       enqueued
       
       -
       
       write
       
       -
       
       timeout 
       
       = 
       
       "30s"
       
[
       
       logging]
       
       # 
       
        The log level ,error、warn、info(
       
        Default )
       
       、debug
       
       level 
       
       = 
       
       "info"
      
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.

Key configuration (cpu/ Memory / disk io/ Overtime / journal )

       
       # 
       
        Disable reporting , The default is  
       
       false
       
       reporting
       
       -
       
       disabled 
       
       = 
       
       true
       
[
       
       meta]
       
       # 
       
        Metadata storage directory 
       
       dir 
       
       = 
       
       "/var/lib/influxdb/meta"
       
[
       
       data]
       
       # 
       
        Data directory 
       
       dir 
       
       = 
       
       "/var/lib/influxdb/data"
       
       # 
       
        Pre - write log directory （write 
       
       ahead 
       
       log）
       
       wal
       
       -
       
       dir 
       
       = 
       
       "/var/lib/influxdb/wal"
       
       # 
       
       fsync 
       
        Write wait time before call , The default is  
       
       0
       
       s,SSD 
       
        Set to  
       
       0
       
       s, Not  
       
       SSD 
       
        Recommended setting as  
       
       0
       
       ms
       
       -
       
       100
       
       ms
       
       # 
       
        To lower the disk  
       
       io, Then increase it appropriately 
       
       wal
       
       -
       
       fsync
       
       -
       
       delay 
       
       = 
       
       "20ms"
       
       # 
       
       inmem 
       
        Memory index （ Need a lot of memory ）,tsi1 
       
        Disk based timing index , The default is  
       
       inmem
       
       index
       
       -
       
       version 
       
       = 
       
       "tsi1"
       
       # 
       
        Query log switch , Turn off as appropriate 
       
       query
       
       -
       
       log
       
       -
       
       enabled 
       
       = 
       
       true
       
       # 
       
        Fragment cache in reject  
       
       /
       
       write 
       
        The maximum memory size that can be written before the request , The default is  
       
       1
       
       g
       
       # 
       
        If the server memory is small , It should be adjusted down properly ; To lower the disk  
       
       io, Then increase it appropriately 
       
       cache
       
       -
       
       max
       
       -
       
       memory
       
       -
       
       size 
       
       = 
       
       "1g"
       
       # 
       
       tsm 
       
        The engine caches the snapshot and writes it to  
       
       tsm 
       
        Cache size of the file , Free memory when finished , The default is  
       
       25
       
       m
       
       # 
       
        If the server memory is small , It should be adjusted down properly ; To lower the disk  
       
       io, Then increase it appropriately 
       
       cache
       
       -
       
       snapshot
       
       -
       
       memory
       
       -
       
       size 
       
       = 
       
       "25m"
       
       # 
       
        After a while , If the partition does not receive write or delete ,tsm 
       
        The engine will take a snapshot of the cache and write it to the new  
       
       TSM 
       
        file , The default is  
       
       10
       
       m
       
       # 
       
        To reduce  
       
       cpu 
       
        Compress computing and disk  
       
       io, Then increase it appropriately , But avoid using memory for too long 
       
       cache
       
       -
       
       snapshot
       
       -
       
       write
       
       -
       
       cold
       
       -
       
       duration 
       
       = 
       
       "10m"
       
       # 
       
        After a while , If  
       
       tsm 
       
        The engine did not receive a write or delete ,tsm 
       
        The engine will compress all in the shard  
       
       TSM 
       
        file , The default is  
       
       4
       
       h
       
       # 
       
        To reduce  
       
       cpu 
       
        Compress computing and disk  
       
       io, Then increase it appropriately , for example  
       
       24
       
       h
       
       compact
       
       -
       
       full
       
       -
       
       write
       
       -
       
       cold
       
       -
       
       duration 
       
       = 
       
       "24h"
       
       # 
       
       tsm 
       
        Compress the maximum amount of data written to the disk per second , The default is  
       
       48
       
       m
       
       # 
       
        To lower the disk  
       
       io, Then increase it appropriately 
       
       compact
       
       -
       
       throughput 
       
       = 
       
       "48m"
       
       # 
       
       tsm 
       
        Compress the maximum amount of data written to disk per second during a short burst , The default is  
       
       48
       
       m
       
       # 
       
        To lower the disk  
       
       io, Then increase it appropriately 
       
       compact
       
       -
       
       throughput
       
       -
       
       burst 
       
       = 
       
       "48m"
       
       # 
       
        Maximum concurrent compression , The default value is  
       
       0 
       
        Will  
       
       50
       
       % 
       
        Of  
       
       CPU 
       
        The number of cores is used to compress 
       
       # 
       
        To reduce  
       
       cpu 
       
        Compress computing and disk  
       
       io, It should be adjusted down properly , But avoid too small to cause the compression cycle to become longer 
       
       max
       
       -
       
       concurrent
       
       -
       
       compactions 
       
       = 
       
       0
       
       # 
       
       tsi1 
       
        Set item ： Prewrite log  
       
       wal 
       
        Threshold size when the file will be compressed into an indexed file 
       
       # 
       
        Smaller thresholds will result in faster compression of log files , And result in lower heap memory usage , But at the cost of write throughput 
       
       # 
       
        Larger thresholds will not compress frequently , Store more... In memory  
       
       series, And provide higher write throughput 
       
       # 
       
        To reduce  
       
       cpu 
       
        Compress computing and disk  
       
       io, Then increase it appropriately , But avoid being too large to occupy memory all the time 
       
       # 
       
        If the server memory is small , It should be adjusted down or kept unchanged 
       
       max
       
       -
       
       index
       
       -
       
       log
       
       -
       
       file
       
       -
       
       size 
       
       = 
       
       "1m"
       
       # 
       
       tsi1 
       
        Set item ：tsi 
       
        The index is used to store the previously calculated  
       
       series 
       
        The internal cache size of the result 
       
       # 
       
        Cache results will be quickly returned from the cache , Without matching  
       
       tag 
       
        Subsequent queries of key values are recalculated 
       
       # 
       
        Set this value to  
       
       0 
       
        Caching will be disabled , This can cause query performance problems 
       
       # 
       
        If the server memory is small , It should be adjusted down properly 
       
       series
       
       -
       
       id
       
       -
       
       set
       
       -
       
       cache
       
       -
       
       size 
       
       = 
       
       100
       
[
       
       coordinator]
       
       # 
       
        Write request timeout , The default is  
       
       10
       
       s
       
       write
       
       -
       
       timeout 
       
       = 
       
       "20s"
       
[
       
       monitor]
       
       # 
       
        Internal record statistics switch , If there is a large amount of data in the production environment , Will affect memory and cpu, Proposed closure 
       
       store
       
       -
       
       enabled 
       
       = 
       
       true
       
       # 
       
        Time interval for recording statistics , If there is a large amount of data in the production environment , But monitoring needs to be turned on , Then increase the interval 
       
       store
       
       -
       
       interval 
       
       = 
       
       "10s"
       
[
       
       http]
       
       # 
       
       http 
       
        Request log , Turn off as appropriate 
       
       log
       
       -
       
       enabled 
       
       = 
       
       true
       
[
       
       logging]
       
       # 
       
        The log level ,error、warn、info(
       
        Default )
       
       、debug, Adjust the grade as appropriate 
       
       level 
       
       = 
       
       "info"
      
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.