telegraf/plugins/outputs/bigquery
gkatzioura 5085f595ac
Bigquery output Plugin (#8634)
2021-03-12 12:16:45 -05:00
..
README.md Bigquery output Plugin (#8634) 2021-03-12 12:16:45 -05:00
bigquery.go Bigquery output Plugin (#8634) 2021-03-12 12:16:45 -05:00
bigquery_test.go Bigquery output Plugin (#8634) 2021-03-12 12:16:45 -05:00

README.md

BigQuery Google Cloud Output Plugin

This plugin writes to the [Google Cloud BigQuery][bigquery] and requires [authentication][] with Google Cloud using either a service account or user credentials

This plugin accesses APIs which are [chargeable][pricing]; you might incur costs.

Requires project to specify where BigQuery entries will be persisted.

Requires dataset to specify under which BigQuery dataset the corresponding metrics tables reside.

Each metric should have a corresponding table to BigQuery. The schema of the table on BigQuery:

  • Should contain the field timestamp which is the timestamp of a telegraph metrics
  • Should contain the metric's tags with the same name and the column type should be set to string.
  • Should contain the metric's fields with the same name and the column type should match the field type.

Configuration

[[outputs.bigquery]]
  ## GCP Project
  project = "erudite-bloom-151019"

  ## The BigQuery dataset
  dataset = "telegraf"

  ## Timeout for BigQuery operations.
  # timeout = "5s"

  ## Character to replace hyphens on Metric name
  # replace_hyphen_to = "_"

Restrictions

Avoid hyphens on BigQuery tables, underlying SDK cannot handle streaming inserts to Table with hyphens.

In cases of metrics with hyphens please use the Rename Processor Plugin.

In case of a metric with hyphen by default hyphens shall be replaced with underscores (_). This can be altered using the replace_hyphen_to configuration property.

Available data type options are:

  • integer
  • float or long
  • string
  • boolean

All field naming restrictions that apply to BigQuery should apply to the measurements to be imported.

Tables on BigQuery should be created beforehand and they are not created during persistence

Pay attention to the column timestamp since it is reserved upfront and cannot change. If partitioning is required make sure it is applied beforehand.