Skip to content

Latest commit

 

History

History
83 lines (65 loc) · 2.24 KB

INFLUXDB.md

File metadata and controls

83 lines (65 loc) · 2.24 KB
Raw

InfluxDB

Installation with Docker

To run Influxdb (with chronograf + telegraf) as a docker services:

docker-compose -f influxdb/docker-compose.yml up -d

Design & Schema

Each point consists of

  • time
  • measurement
    • e.g. ais_sequence
  • tags: containing metadata
    • e.g. mmsi=999464786
    • e.g. sequence_id=999464786#9823492341
  • fields (key-value)
    • e.g. longitude=551464.14, latitude=124411.14

Why we choose mmsi as a tag, and longitude, latitude as fields?

Because of:

  • “the lower the cardinality, the more duplicated elements in a column”
  • mmsi has low cardinality
  • Longitude and latitude have high cardinality and cannot be chosen to be tags, that would results to large number of series ⇒ less efficient

Can we put all attributes as fields?

  • Yes but very bad, why?
    • Tags are indexed while fields are not indexed
    • If we make them all fields, then it will be hard to pull up them all out from the database.

Ingesting CSV file

Many ways:

  • Custom CSV parser and submit with curl command (working but not suffecient)
  • csv-to-influx library (working with bugs and errors)
  • Telegraf (try but to success)
  • Logstash (try but to success)

Querying InfluxDB

Refer to influxdb website to see the full documentation

Use docker command to access influxdb container and then influx to access influxdb shell CLI, :

docker exec -it my-influxdb /bin/bash

root@my-influxdb:/# influx

Connected to http://localhost:8086 version 1.7.4
InfluxDB shell version: 1.7.4
Enter an InfluxQL query
>
  • Sample query
select latitude,longitude 
from AIS20190204 
where time >= '2019-02-04T19:30:00Z' and time <= '2019-02-04T20:00:00Z' 
group by mmsi limit 5

Chronograf

If you have already ingested data to influxdb and you want to view it with Chronograf, you can do that by accessing UI with http://localhost:8888:

Sample query:

select *
from AIS.autogen.ais_me 
where time >= '2019-02-04T19:30:00Z' and time <= '2019-02-04T20:00:00Z' and mmsi = '212746000'
group by mmsi

Or

select count(*)
from AIS.autogen.ais_me