InfluxDB 1.x

The contents on this page is specific to InfluxDB 1.x. For information on 2.x, see the main InfluxDB article.

Using the InfluxDB command line

Get started with the InfluxDB command line. If using a docker container, you can connect to localhost without credentials on the default port. You can execute queries either by entering it to the influx shell or by piping it to influx.

# echo "show databases" | influx -host localhost

# influx -host localhost
Connected to http://localhost:8086 version 1.7.10
InfluxDB shell version: 1.7.10
> show databases

Cheat Sheet

Action	Command
Show databases	show databases;
Create a database	create database $database;
Select a database	use $database;
Show all measurements	show measurements
Drop a measurement	drop measurement weather;
Show measurements matching a tag	show measurement where tag = 'value' show measurement where tag =~ /regex.*/
Show running queries	show queries;
Show all users	show users;
Show retention policies on a database	show retention policies on "$database"
Show all series	show series from temperature
Show all tags	show tag keys from "$measurement"
Copy data to another measurement/database	select * INTO newdb.autogen.cpu FROM olddb.autogen.cpu WHERE host = 'arc' GROUP BY *
Continuous Queries	show continuous queries
Delete a continuous query	drop continuous query cq-name ON measurement

Deleting Measurement

To delete an entire measurement, you must enter the influx command line, set the database and then run DROP MEASUREMENT. You cannot do this from Chronograf.

# influx -host localhost
Connected to http://localhost:8086 version 1.7.10
InfluxDB shell version: 1.7.10
> show databases;
nodered
> use nodered
Using database nodered
> drop measurement local_weather;
>

Delete a Series

To delete a specific series based on tag values, use the DELETE FROM query from the influx command line.

# influx -host localhost
Connected to http://localhost:8086 version 1.7.10
InfluxDB shell version: 1.7.10
> use nodered
Using database nodered
> DELETE FROM "power" WHERE device = 'AWP04L-1'

Deleting a series will not drop any fields even if the field no longer has any values. See the next section on dropping a field.

Copy measurements to another database

You can copy a measurement to another database by using the SELECT * INTO type query. This might be desirable if you wish to make a backup of the data for only a specific set of measurements based on a tag. For example, to copy data from the source database olddb to the destination database newdb, but only for measurements where the tag host is arc:

SELECT * INTO newdb.autogen.cpu
FROM olddb.autogen.cpu
WHERE host = 'arc'
GROUP BY *

Down-sampling Data

See: https://docs.influxdata.com/influxdb/v1.8/guides/downsample_and_retain/

You need to have a separate measurement for the down sampled data. Use a continuous query (CQ) to push data from the non-down sampled measurement into the down sampled measurement.

Eg:

CREATE RETENTION POLICY "longterm" ON telegraf DURATION 1825d REPLICATION 1

CREATE CONTINUOUS QUERY "30m_telegraf_cpu" ON telegraf 
BEGIN 
SELECT 
	mean(usage_guest) as usage_guest,
	mean(usage_guest_nice) as usage_guest_nice,
	mean(usage_idle) as usage_idle,
	mean(usage_iowait) as usage_iowait,
	mean(usage_irq) as usage_irq,
	mean(usage_nice) as usage_nice,
	mean(usage_softirq) as usage_softirq,
	mean(usage_steal) as usage_steal,
	mean(usage_system) as usage_system,
	mean(usage_user) as usage_user
INTO telegraf.longterm.:measurement 
FROM telegraf.autogen.cpu GROUP BY time(30m)
END

Drop a Field

There is no capability to drop a field in InfluxDB. There is a feature request for it at https://github.com/influxdata/influxdb/issues/6150.

A work around is to SELECT the fields that you want into a placeholder measurement, drop the original measurement with the extraneous fields, and then re-select everything from the placeholder back into the original measurement.

# influx -host localhost
Connected to http://localhost:8086 version 1.7.10
InfluxDB shell version: 1.7.10
> use nodered
Using database nodered

## Note the extra field starting with an uppercase
> select * FROM power LIMIT 1;
name: power
time                Amperage ApparentPower Factor ReactivePower Voltage Wattage amperage apparentPower device               factor reactivePower voltage wattage
----                -------- ------------- ------ ------------- ------- ------- -------- ------------- ------               ------ ------------- ------- -------
1588742430632718326                                                             0.033                  02200194dc4f22137d7f

## Copy desired fields to a placeholder measurement
> select time, amperage, apparentPower, device, factor, reactivePower, voltage, wattage INTO demo FROM power group by *
name: result
time written
---- -------
0    574590

## Drop the original measurement
> drop measurement power

## Re-insert everything back into the original measurement
> select * INTO power FROM demo
name: result
time written
---- -------
0    574590

## Drop the placeholder
> drop measurement demo

Because you are re-copying the data again, this might not be a good solution for extremely large measurements.

Show unique tag values

To show all tag values that are in a measurement (such as the 'jobs' measurement in the example below), use the SHOW TAG VALUES query. More information can be found from the InfluxDB documentation.

# influx -host localhost
Connected to http://localhost:8086 version 1.7.10
InfluxDB shell version: 1.7.10
> use x
Using database x

> show tag values from jobs with key = partition
name: jobs                                      
key       value                                 
---       -----                                 
partition apophis                               
partition apophis-bf                            
partition bigmem                                
partition breezy                                
partition cpu2013                               
partition cpu2019                               
partition gpu                                   
partition gpu-v100                              
partition lattice                               
partition parallel

In this case, the measurement 'jobs' has a tag 'partition' which has these distinct values. To exclude some value in this query, add a WHERE clause that excludes certain values. Eg.:

> SHOW TAG VALUES FROM jobs WITH KEY = partition WHERE partition !~ /bigmem/

This type of query is great when generating dashboard variables in Grafana.

Writing data

The schema of a InfluxDB database is based on the very first initial data types added. Adding additional data with a different type will be rejected.

• https://docs.influxdata.com/influxdb/v1.7/guides/writing_data/
• https://www.influxdata.com/blog/how-to-send-sensor-data-to-influxdb-from-an-arduino-uno/

With Curl using the line protocol

Using curl, you can POST values to a particular database using the line protocol. For more information, see https://docs.influxdata.com/influxdb/v0.9/write_protocols/line/

$ curl -X POST 'http://influxdb/write?db=lsf' --data-binary 'measurement,tag1=x,tag2=x recorded_value1=x `date +%s"000000000"`'

For multiple entries, write multiple lines separated by a \n newline and POST with Curl like above.

$ cat /tmp/measurements | curl -X POST --data-binary @- 'http://influxdb/write?db=database'

Large inputs might result in a 413 Payload Too Large error and requires you to split the input. The simplest way is to just split the input. Eg. sending data in batches of 500 lines:

$ cat /tmp/measurements | split -l 500 --filter="curl -X POST --data-binary @- 'http://influxdb/write?db=database'"

Backing up data

Influxd

With Influxd, to backup:

# influxd backup -portable -database telegraftwo /tmp/backup

To restore, copy the generated files to the host you want to restore to and run:

# influxd restore -portable -database telegraftwo /tmp/backup

Curl

There is no easy way to restore the JSON output that's generated using this method. However, it's in a readable format for other purposes.

$ curl "http://influxdb/query?db=slurm-metrics&" --data-urlencode "q=SELECT * FROM $series;" > series

Migrating from Prometheus

There is a go utility called prom2influx (https://github.com/zhyon404/prom2influx). More information documented at: https://docs.tibco.com/pub/ftl/5.4.0/doc/html/GUID-BAE2C630-90B2-47FC-A7D6-97AB16065A0C.html

To migrate data from Prometheus to InfluxDB, use the prom2influx migration utility

Ensure that the Prometheus server is running.

Ensure that the InfluxDB server is running.

Run the migration utility. For example, this command migrates all the data points in Prometheus:

prom2influx -prometheus http://prom_host:9090 -influx http://influx_host:8086

The duration of this step depends on the amount of data you migrate, network communication bandwidth and speeds, and other factors. You can migrate a subset of the data by specifying the -start and -range parameters, for example:

prom2influx -prometheus http://prom_host:9090 -influx http://influx_host:8086 -start 2018-03-12T12:00:57-05:00 -range 1y

Dealing with non-periodic measurements with the sum aggregate

For certain measurements, a sum aggregate needs to be used. However, this requires measurements to occur on a specific time interval or else the summed measurement would be incorrect. In a sense, what we really need is to sum the measurements on a small time interval and then average the results.

This can be accomplished using nested queries. One of the metrics I wanted to graph which was logged every 24 hours used:

SELECT avg("x") FROM 
(SELECT sum("usage") as x FROM "storage_usage" WHERE "partition" = 'work'  AND $timeFilter  GROUP BY time(1h)  )   
GROUP BY time($__interval) fill(null)

Measurements taken more frequently than 24 hours but less frequent than 1 hour won't cause the result to be incorrect.