influxdb-ruby

The official Ruby client library for InfluxDB. Maintained by @toddboom and @dmke.

Platform support

Support for InfluxDB v0.8.x is now deprecated. The final version of this library that will support the older InfluxDB interface is v0.1.9, which is available as a gem and tagged on this repository.

If you're reading this message, then you should only expect support for InfluxDB v0.9.1 and higher.

Ruby support

Since v0.4.0, this gem requires Ruby >= 2.2.0. Support for MRI < 2.2 is still available in the v0.3.x series, see stable-03 branch for documentation.

Installation

$ [sudo] gem install influxdb

Or add it to your Gemfile, and run bundle install.

Usage

All examples assume you have a require "influxdb" in your code.

Creating a client

Connecting to a single host:

influxdb = InfluxDB::Client.new  # default connects to localhost:8086

# or
influxdb = InfluxDB::Client.new host: "influxdb.domain.com"

Connecting to multiple hosts (with built-in load balancing and failover):

influxdb = InfluxDB::Client.new hosts: ["influxdb1.domain.com", "influxdb2.domain.com"]

Using a configuration URL

You can also provide a URL to connect to your server. This is particulary useful for 12-factor apps, i.e. you can put the configuration in an environment variable:

url = ENV["INFLUXDB_URL"] || "https://influxdb.example.com:8086/database_name?retry=3"
influxdb = InfluxDB::Client.new url: url

Please note, that the config options found in the URL have a lower precedence than those explicitly given in the options hash. This means, that the following sample will use an open-timeout of 10 seconds:

url = "https://influxdb.example.com:8086/database_name?open_timeout=3"
influxdb = InfluxDB::Client.new url: url, open_timeout: 10

Writing data

Write some data:

username = 'foo'
password = 'bar'
database = 'site_development'
name     = 'foobar'

influxdb = InfluxDB::Client.new database, username: username, password: password

# Enumerator that emits a sine wave
Value = (0..360).to_a.map {|i| Math.send(:sin, i / 10.0) * 10 }.each

loop do
  data = {
    values: { value: Value.next },
    tags:   { wave: 'sine' } # tags are optional
  }

  influxdb.write_point(name, data)

  sleep 1
end

Write data with time precision (precision can be set in 2 ways):

username       = 'foo'
password       = 'bar'
database       = 'site_development'
name           = 'foobar'
time_precision = 's'

# either in the client initialization:
influxdb = InfluxDB::Client.new database,
  username: username,
  password: password,
  time_precision: time_precision

data = {
  values: { value: 0 },
  timestamp: Time.now.to_i # timestamp is optional, if not provided point will be saved with current time
}

influxdb.write_point(name, data)

# or in a method call:
influxdb.write_point(name, data, time_precision)

A Note About Time Precision

The default precision (in this Rubygem) is "s" (second), as Ruby's Time#to_i operates on this resolution.

If you write data points with sub-second resolution (for example via Time#to_r), you need to configure your client instance with an appropiate time_precision option and you'll need to provide a timestamp value which reflects this precision:
influxdb = InfluxDB::Client.new time_precision: "ms"
time = (Time.now.to_r * 10**6).to_i
influxdb.write_point("foobar", { values: { n: 42 }, timestamp: time })

Allowed values for time_precision are:

"ns" or nil for nanosecond
"u" for microsecond
"ms" for millisecond
"s" for second
"m" for minute
"h" for hour

Write data with a specific retention policy:

database  = 'site_development'
name      = 'foobar'
precision = 's'
retention = '1h.cpu'

influxdb = InfluxDB::Client.new database,
  username: "foo",
  password: "bar"

data = {
  values:    { value: 0 },
  tags:      { foo: 'bar', bar: 'baz' },
  timestamp: Time.now.to_i
}

influxdb.write_point(name, data, precision, retention)

Write data while choosing the database:

database  = 'site_development'
name      = 'foobar'
precision = 's'
retention = '1h.cpu'

influxdb = InfluxDB::Client.new {
  username: "foo",
  password: "bar"
}

data = {
  values:    { value: 0 },
  tags:      { foo: 'bar', bar: 'baz' }
  timestamp: Time.now.to_i
}

influxdb.write_point(name, data, precision, retention, database)

Write multiple points in a batch (performance boost):

data = [
  {
    series: 'cpu',
    tags:   { host: 'server_1', region: 'us' },
    values: { internal: 5, external: 0.453345 }
  },
  {
    series: 'gpu',
    values: { value: 0.9999 },
  }
]

influxdb.write_points(data)

# you can also specify precision in method call

precision = 'm'
influxdb.write_points(data, precision)

Write multiple points in a batch with a specific retention policy:

data = [
  {
    series: 'cpu',
    tags:   { host: 'server_1', region: 'us' },
    values: { internal: 5, external: 0.453345 }
  },
  {
    series: 'gpu',
    values: { value: 0.9999 },
  }
]

precision = 'm'
retention = '1h.cpu'
influxdb.write_points(data, precision, retention)

Write asynchronously (note that a retention policy cannot be specified for asynchronous writes):

database = 'site_development'
name     = 'foobar'

influxdb = InfluxDB::Client.new database,
  username: "foo",
  password: "bar",
  async:    true

data = {
  values:    { value: 0 },
  tags:      { foo: 'bar', bar: 'baz' },
  timestamp: Time.now.to_i
}

influxdb.write_point(name, data)

Using async: true is a shortcut for the following:

async_options = {
  # number of points to write to the server at once
  max_post_points:    1000,
  # queue capacity
  max_queue_size:     10_000,
  # number of threads
  num_worker_threads: 3,
  # max. time (in seconds) a thread sleeps before
  # checking if there are new jobs in the queue
  sleep_interval:     5
}

influxdb = InfluxDB::Client.new database, async: async_options

Write data via UDP (note that a retention policy cannot be specified for UDP writes):

influxdb = InfluxDB::Client.new udp: { host: "127.0.0.1", port: 4444 }

name = 'hitchhiker'

data = {
  values: { value: 666 },
  tags:   { foo: 'bar', bar: 'baz' }
}

influxdb.write_point(name, data)

Discard write errors:

influxdb = InfluxDB::Client.new(
  udp: { host: "127.0.0.1", port: 4444 },
  discard_write_errors: true
)

influxdb.write_point('hitchhiker', { values: { value: 666 } })

Querying

database = 'site_development'
influxdb = InfluxDB::Client.new database,
  username: "foo",
  password: "bar"

# without a block:
influxdb.query 'select * from time_series_1 group by region'

# results are grouped by name, but also their tags:
#
# [
#   {
#     "name"=>"time_series_1",
#     "tags"=>{"region"=>"uk"},
#     "values"=>[
#       {"time"=>"2015-07-09T09:03:31Z", "count"=>32, "value"=>0.9673},
#       {"time"=>"2015-07-09T09:03:49Z", "count"=>122, "value"=>0.4444}
#     ]
#   },
#   {
#     "name"=>"time_series_1",
#     "tags"=>{"region"=>"us"},
#     "values"=>[
#       {"time"=>"2015-07-09T09:02:54Z", "count"=>55, "value"=>0.4343}
#     ]
#   }
# ]

# with a block:
influxdb.query 'select * from time_series_1 group by region' do |name, tags, points|
  printf "%s [ %p ]\n", name, tags
  points.each do |pt|
    printf "  -> %p\n", pt
  end
end

# result:
# time_series_1 [ {"region"=>"uk"} ]
#   -> {"time"=>"2015-07-09T09:03:31Z", "count"=>32, "value"=>0.9673}
#   -> {"time"=>"2015-07-09T09:03:49Z", "count"=>122, "value"=>0.4444}]
# time_series_1 [ {"region"=>"us"} ]
#   -> {"time"=>"2015-07-09T09:02:54Z", "count"=>55, "value"=>0.4343}

If you would rather receive points with integer timestamp, it's possible to set epoch parameter:

# globally, on client initialization:
influxdb = InfluxDB::Client.new database, epoch: 's'

influxdb.query 'select * from time_series group by region'
# [
#   {
#     "name"=>"time_series",
#     "tags"=>{"region"=>"uk"},
#     "values"=>[
#       {"time"=>1438411376, "count"=>32, "value"=>0.9673}
#     ]
#   }
# ]

# or for a specific query call:
influxdb.query 'select * from time_series group by region', epoch: 'ms'
# [
#   {
#     "name"=>"time_series",
#     "tags"=>{"region"=>"uk"},
#     "values"=>[
#       {"time"=>1438411376000, "count"=>32, "value"=>0.9673}
#     ]
#   }
# ]

Working with parameterized query strings works as expected:

influxdb = InfluxDB::Client.new database

named_parameter_query = "select * from time_series_0 where time > %{min_time}"
influxdb.query named_parameter_query, params: { min_time: 0 }
# compiles to:
#   select * from time_series_0 where time > 0

positional_params_query = "select * from time_series_0 where f = %{1} and i < %{2}"
influxdb.query positional_params_query, params: ["foobar", 42]
# compiles to (note the automatic escaping):
#   select * from time_series_0 where f = 'foobar' and i < 42

Advanced Topics

Administrative tasks

Create a database:

database = 'site_development'

influxdb.create_database(database)

Delete a database:

database = 'site_development'

influxdb.delete_database(database)

List databases:

influxdb.list_databases

Create a user for a database:

database = 'site_development'
new_username = 'foo'
new_password = 'bar'
permission = :write

# with all permissions
influxdb.create_database_user(database, new_username, new_password)

# with specified permission - options are: :read, :write, :all
influxdb.create_database_user(database, new_username, new_password, permissions: permission)

Update a user password:

username = 'foo'
new_password = 'bar'

influxdb.update_user_password(username, new_password)

Grant user privileges on database:

username = 'foobar'
database = 'foo'
permission = :read # options are :read, :write, :all

influxdb.grant_user_privileges(username, database, permission)

Revoke user privileges from database:

username = 'foobar'
database = 'foo'
permission = :write # options are :read, :write, :all

influxdb.revoke_user_privileges(username, database, permission)

Delete a user:

username = 'foobar'

influxdb.delete_user(username)

List users:

influxdb.list_users

Create cluster admin:

username = 'foobar'
password = 'pwd'

influxdb.create_cluster_admin(username, password)

List cluster admins:

influxdb.list_cluster_admins

Revoke cluster admin privileges from user:

username = 'foobar'

influxdb.revoke_cluster_admin_privileges(username)

Continuous Queries

List continuous queries of a database:

database = 'foo'

influxdb.list_continuous_queries(database)

Create a continuous query for a database:

database = 'foo'
name = 'clicks_count'
query = 'SELECT COUNT(name) INTO clicksCount_1h FROM clicks GROUP BY time(1h)'

influxdb.create_continuous_query(name, database, query)

Additionally, you can specify the resample interval and the time range over which the CQ runs:

influxdb.create_continuous_query(name, database, query, resample_every: "10m", resample_for: "65m")

Delete a continuous query from a database:

database = 'foo'
name = 'clicks_count'

influxdb.delete_continuous_query(name, database)

Retention Policies

List retention policies of a database:

database = 'foo'

influxdb.list_retention_policies(database)

Create a retention policy for a database:

database    = 'foo'
name        = '1h.cpu'
duration    = '10m'
replication = 2

influxdb.create_retention_policy(name, database, duration, replication)

Delete a retention policy from a database:

database = 'foo'
name     = '1h.cpu'

influxdb.delete_retention_policy(name, database)

Alter a retention policy for a database:

database    = 'foo'
name        = '1h.cpu'
duration    = '10m'
replication = 2

influxdb.alter_retention_policy(name, database, duration, replication)

Reading data

(De-) Normalization

By default, InfluxDB::Client will denormalize points (received from InfluxDB as columns and rows). If you want to get raw data add denormalize: false to the initialization options or to query itself:

influxdb.query 'select * from time_series_1 group by region', denormalize: false

# [
#   {
#     "name"=>"time_series_1",
#     "tags"=>{"region"=>"uk"},
#     "columns"=>["time", "count", "value"],
#     "values"=>[
#       ["2015-07-09T09:03:31Z", 32, 0.9673],
#       ["2015-07-09T09:03:49Z", 122, 0.4444]
#     ]
#   },
#   {
#     "name"=>"time_series_1",
#     "tags"=>{"region"=>"us"},
#     "columns"=>["time", "count", "value"],
#     "values"=>[
#       ["2015-07-09T09:02:54Z", 55, 0.4343]
#     ]
#   }
# ]


influxdb.query 'select * from time_series_1 group by region', denormalize: false do |name, tags, points|
  printf "%s [ %p ]\n", name, tags
  points.each do |key, values|
    printf "  %p -> %p\n", key, values
  end
end

# time_series_1 [ {"region"=>"uk"} ]
#   columns -> ["time", "count", "value"]
#   values -> [["2015-07-09T09:03:31Z", 32, 0.9673], ["2015-07-09T09:03:49Z", 122, 0.4444]]}
# time_series_1 [ {"region"=>"us"} ]
#   columns -> ["time", "count", "value"]
#   values -> [["2015-07-09T09:02:54Z", 55, 0.4343]]}

You can also pick the database to query from:

influxdb.query 'select * from time_series_1', database: 'database'

Streaming response

If you expect large quantities of data in a response, you may want to enable JSON streaming by setting a chunk_size:

influxdb = InfluxDB::Client.new database,
  username:   username,
  password:   password,
  chunk_size: 10000

See the official documentation for more details.

Retry

By default, InfluxDB::Client will keep trying (with exponential fall-off) to connect to the database until it gets a connection. If you want to retry only a finite number of times (or disable retries altogether), you can pass the :retry option.

:retry can be either true, false or an Integer to retry infinite times, disable retries or retry a finite number of times, respectively. Passing 0 is equivalent to false and -1 is equivalent to true.

$ irb -r influxdb
> influxdb = InfluxDB::Client.new 'database', retry: 8
=> #<InfluxDB::Client:0x00000002bb5ce0 ...>

> influxdb.query 'select * from serie limit 1'
E, [2016-08-31T23:55:18.287947 #23476] WARN  -- InfluxDB: Failed to contact host localhost: #<Errno::ECONNREFUSED: Failed to open TCP connection to localhost:8086 (Connection refused - connect(2) for "localhost" port 8086)> - retrying in 0.01s.
E, [2016-08-31T23:55:18.298455 #23476] WARN  -- InfluxDB: Failed to contact host localhost: #<Errno::ECONNREFUSED: Failed to open TCP connection to localhost:8086 (Connection refused - connect(2) for "localhost" port 8086)> - retrying in 0.02s.
E, [2016-08-31T23:55:18.319122 #23476] WARN  -- InfluxDB: Failed to contact host localhost: #<Errno::ECONNREFUSED: Failed to open TCP connection to localhost:8086 (Connection refused - connect(2) for "localhost" port 8086)> - retrying in 0.04s.
E, [2016-08-31T23:55:18.359785 #23476] WARN  -- InfluxDB: Failed to contact host localhost: #<Errno::ECONNREFUSED: Failed to open TCP connection to localhost:8086 (Connection refused - connect(2) for "localhost" port 8086)> - retrying in 0.08s.
E, [2016-08-31T23:55:18.440422 #23476] WARN  -- InfluxDB: Failed to contact host localhost: #<Errno::ECONNREFUSED: Failed to open TCP connection to localhost:8086 (Connection refused - connect(2) for "localhost" port 8086)> - retrying in 0.16s.
E, [2016-08-31T23:55:18.600936 #23476] WARN  -- InfluxDB: Failed to contact host localhost: #<Errno::ECONNREFUSED: Failed to open TCP connection to localhost:8086 (Connection refused - connect(2) for "localhost" port 8086)> - retrying in 0.32s.
E, [2016-08-31T23:55:18.921740 #23476] WARN  -- InfluxDB: Failed to contact host localhost: #<Errno::ECONNREFUSED: Failed to open TCP connection to localhost:8086 (Connection refused - connect(2) for "localhost" port 8086)> - retrying in 0.64s.
E, [2016-08-31T23:55:19.562428 #23476] WARN  -- InfluxDB: Failed to contact host localhost: #<Errno::ECONNREFUSED: Failed to open TCP connection to localhost:8086 (Connection refused - connect(2) for "localhost" port 8086)> - retrying in 1.28s.
InfluxDB::ConnectionError: Tried 8 times to reconnect but failed.

List of configuration options

This index might be out of date. Please refer to InfluxDB::DEFAULT_CONFIG_OPTIONS, found in lib/influxdb/config.rb for the source of truth.

Category	Option	Default value	Notes
HTTP connection	`:host` or `:hosts`	"localhost"	can be an array and can include port
	`:port`	8086	fallback port, unless provided by `:host` option
	`:prefix`	""	URL path prefix (e.g. server is behind reverse proxy)
	`:username`	"root"	user credentials
	`:password`	"root"	user credentials
	`:open_timeout`	5	socket timeout
	`:read_timeout`	300	socket timeout
	`:auth_method`	"params"	"params", "basic_auth" or "none"
Retry	`:retry`	-1	max. number of retry attempts (reading and writing)
	`:initial_delay`	0.01	initial wait time (doubles every retry attempt)
	`:max_delay`	30	max. wait time when retrying
SSL/HTTPS	`:use_ssl`	false	whether or not to use SSL (HTTPS)
	`:verify_ssl`	true	verify vertificate when using SSL
	`:ssl_ca_cert`	false	path to or name of CA cert
Database	`:database`	empty	name of database
	`:time_precision`	"s"	time resolution for data send to server
	`:epoch`	false	time resolution for server responses (false = server default)
Writer	`:async`	false	Async options hash, details here
	`:udp`	false	UDP connection info, details here
	`:discard_write_errors`	false	suppress UDP socket errors
Query	`:chunk_size`	empty	details here
	`:denormalize`	true	format of result

Testing

git clone git@github.com:influxdata/influxdb-ruby.git
cd influxdb-ruby
bundle
bundle exec rake

Contributing

Fork this repository on GitHub.
Make your changes.
- Add tests.
- Add an entry in the CHANGELOG.md in the "unreleased" section on top.
Run the tests: bundle exec rake.
Send a pull request.
- Please rebase against the master branch.
If your changes look good, we'll merge them.

Name		Name	Last commit message	Last commit date
Latest commit History 471 Commits
lib		lib
smoke		smoke
spec		spec
.gitignore		.gitignore
.rubocop.yml		.rubocop.yml
.travis.yml		.travis.yml
CHANGELOG.md		CHANGELOG.md
Gemfile		Gemfile
LICENSE.txt		LICENSE.txt
README.md		README.md
Rakefile		Rakefile
influxdb.gemspec		influxdb.gemspec

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

influxdb-ruby

Contents

Platform support

Ruby support

Installation

Usage

Creating a client

Using a configuration URL

Writing data

A Note About Time Precision

Querying

Advanced Topics

Administrative tasks

Continuous Queries

Retention Policies

Reading data

(De-) Normalization

Streaming response

Retry

List of configuration options

Testing

Contributing

About

Releases

Packages

Languages

License

onlynone/influxdb-ruby

Folders and files

Latest commit

History

Repository files navigation

influxdb-ruby

Contents

Platform support

Ruby support

Installation

Usage

Creating a client

Using a configuration URL

Writing data

A Note About Time Precision

Querying

Advanced Topics

Administrative tasks

Continuous Queries

Retention Policies

Reading data

(De-) Normalization

Streaming response

Retry

List of configuration options

Testing

Contributing

About

Resources

License

Stars

Watchers

Forks

Releases

Packages 0

Languages

Packages