Difference between revisions of "InfluxDB"

From air
Jump to navigation Jump to search
 
Line 30: Line 30:
 
<pre>
 
<pre>
 
# for 64-bit systems
 
# for 64-bit systems
wget https://s3.amazonaws.com/influxdb/influxdb_0.9.6.1_amd64.deb
+
wget https://dl.influxdata.com/influxdb/releases/influxdb_0.13.0_amd64.deb
sudo dpkg -i influxdb_0.9.6.1_amd64.deb
+
sudo dpkg -i influxdb_0.13.0_amd64.deb
  +
sudo service influxdb start
  +
sudo service influxdb status
 
</pre>
 
</pre>
   

Latest revision as of 17:46, 1 September 2016

La version courante est la version 0.10

Vers la page de la version 0.8 InfluxDB 0.8


http://influxdb.com/

https://github.com/influxdb/influxdb

open-source distributed time series database for storing metrics, events, and analytics.

developped in Go language

  • retention policy : data is automatically deleted after the retention period
  • sharding :
  • replication :

Voir https://speakerdeck.com/pauldix/introducing-influxdb-an-open-source-distributed-time-series-database

Demo sur http://play.influxdb.com/ (source code)

InfluxDB @ AIR

Installation

http://influxdb.com/download/

On Linux (Debian)

# for 64-bit systems
wget https://dl.influxdata.com/influxdb/releases/influxdb_0.13.0_amd64.deb
sudo dpkg -i influxdb_0.13.0_amd64.deb 
sudo service influxdb start
sudo service influxdb status

On OS X

brew update
brew install influxdb

With Docker

See https://hub.docker.com/r/tutum/influxdb/

Démarrage

On Linux Debian

Configurer Influxdb si besoin en éditant /etc/influxdb/influxdb.conf

sudo service influxdb start
ps wwwax | grep influxdb

L'arrêt se fait avec

sudo service influxdb stop

On MacOS X

sudo /etc/init.d/influxdb start

Naviguez sur http://localhost:8083/ username : root & password : root

Pensez à changer ces valeurs dans /usr/local/etc/influxdb.conf quand vous mettez influxdb en production

Le fichier de configuration peut être généré au moyen de :

/opt/influxdb/influxd config > /etc/influxdb/influxdb.generated.conf

Remarque: les interfaces pour UDP, Graphite, ... peuvent être activées.

Remarque: en production, il convient de sécuriser les ports (ie HTTPS, ...) si le serveur n'est pas derrière un firewall.

Schema Design

Le schéma de la base est constituée de time series contenant des datapoints.

Un datapoint est un n-uplets:

timeserie,tag=value,tag=value, measure=value,measure=value timestamp

Les datapoints sont indexés sur le timestamp (primaire) et sur les tags (secondaires)

Premiers Pas

Via le shell

influx
influx -host localhost

Créer la base

CREATE DATABASE mydb
SHOW DATABASES
USE mydb

Insérer des points

INSERT cpu,host=serverA,region=us_west value=0.64
INSERT cpu,host=serverA,region=us_east value=0.70
INSERT cpu,host=serverA,region=us_west value=0.70
INSERT cpu,host=serverA,region=us_east value=0.90

INSERT payment,device=mobile,product=Notepad,method=credit billed=33,licenses=3i 1434067467100293230

INSERT stock,symbol=AAPL bid=127.46,ask=127.48

INSERT temperature,machine=unit42,type=assembly external=25,internal=37 1434067467000000000

Consulter la base

show series

SELECT * FROM cpu
SELECT * FROM cpu WHERE value >= 0.7
SELECT * FROM cpu LIMIT 1
SELECT value from cpu WHERE time > now() - 7d
SELECT mean(value) FROM cpu WHERE time > now() - 1h GROUP BY time(10m);

SELECT * FROM temperature

Consulter la base en continue (Continuous Queries)

TODO

# Create a CQ to sample the 95% value from 5 minute buckets of the response_times measurement
CREATE CONTINUOUS QUERY response_times_percentile ON mydb BEGIN
  SELECT percentile(value, 95) INTO "response_times.percentiles.5m.95" FROM response_times GROUP BY time(5m)
END


Lister et supprimer des requêtes en continu

SHOW CONTINUOUS QUERIES

DROP CONTINUOUS QUERY response_times_percentile ON mydb

Changer la rétention de la base mydb (ie les données ne sont conservées qu'un jour)

ALTER RETENTION POLICY monitor ON mydb DURATION 1d REPLICATION 1 DEFAULT


Consulter les statistiques

SHOW STATS


Consulter le diagnostique

SHOW DIAGNOSTICS

Consulter le cluster

SHOW SERVERS


Consulter la base interne _internal

USE _internal
show measurements
show series limit 10
select * from httpd limit 10
select * from runtime limit 10

Via l'interface REST

curl -G 'http://localhost:8086/query?pretty=true' --data-urlencode "db=mydb" --data-urlencode \
 "q=SELECT value FROM cpu WHERE region='us_west'"

Requêtes multiples

curl -G 'http://localhost:8086/query?pretty=true' --data-urlencode "db=mydb" --data-urlencode \
 "q=SELECT value FROM cpu WHERE region='us_west';SELECT count(value) FROM cpu WHERE region='us_west'"

Affichage en seconde

curl -G 'http://localhost:8086/query?pretty=true' --data-urlencode "db=mydb" \
 --data-urlencode "epoch=s" --data-urlencode "q=SELECT value FROM cpu WHERE region='us_west'"

Affichage d'un 'chunk' de 20000

curl
curl -G 'http://localhost:8086/query?pretty=true' --data-urlencode "db=mydb" \
 --data-urlencode "chunk_size=20000" --data-urlencode "q=SELECT * FROM cpu"

Via l'interface Web

Depuis l'interface web http://localhost:8083/, Créez une base : mydb

Clustering & Replication

Influxdb utilise le protocole Raft pour l'élection d'un leader : il faut configurer au minimum 3 machines avec des noeuds meta. Remarque : une machine pour être un noeud meta seul, un noeud data seul, ou bien les deux. Le nombre de noeuds data dépend du niveau de réplication désiré.


https://docs.influxdata.com/influxdb/v0.10/guides/clustering/

Vérifier la configuration sous le shell influx

SHOW SERVERS

Backup

Manuel

influxd backup -database telegraf -retention default -since 2016-02-01T00:00:00Z /tmp/backup/telegraf-2016-02-01.db.backup

Restore

sudo service influxdb stop
influxd restore -database telegraf -datadir /var/lib/influxdb/data /tmp/backup/telegraf-2016-02-01.db.backup
sudo service influxdb start
influx -execute 'show databases'

Projets connexes

Telegraf

https://github.com/influxdb/telegraf agent written in Go for collecting metrics from the system it's running on, or from other services, and writing them into InfluxDB.

Chronograf

Chronograf is a single binary web application that you can deploy behind your firewall to do ad hoc exploration of your time series data in InfluxDB. (lien)

Installation et utilisation détaillées ici.

Kapacitor

Kapacitor is a data processing engine. It can process both stream (subscribe realtime) and batch (bulk query) data from InfluxDB. Kapacitor lets you define custom logic to process alerts with dynamic thresholds, match metrics for patterns, compute statistical anomalies, etc. (lien)

Extra