Global configs enable you to fine-tune how dbt runs projects on your machine—whether your personal laptop, an orchestration tool running remotely, or (in some cases) dbt Cloud. They differ from project configs and resource configs, which tell dbt what to run.

Global configs control things like the visual output of logs, the manner in which dbt parses your project, and what to do when dbt finds a version mismatch or a failing model.

These configs are "global" because they are available for all dbt commands, and because they apply across all projects run on the same machine.

Starting in v1.0, you can set global configs in three places. When all three are set, command line flags take precedence, then environment variables, and last profile configs.

Command line (CLI) flags immediately follow dbt and precede your subcommand. When set, CLI flags override environment variables and profile configs.

Use this non-boolean config structure, replacing <THIS-CONFIG> with the config you are enabling or disabling, <SETTING> with the new setting for the config, and <SUBCOMMAND> with the command this config applies to:

$ --<THIS-CONFIG>=<SETTING> <SUBCOMMAND>

Non-boolean config examples:

$ dbt --printer-width=80 run
$ dbt --indirect-selection=eager test

To turn on boolean configs, you would use the --<THIS-CONFIG> CLI flag, and a --no-<THIS-CONFIG> CLI flag to turn off boolean configs, replacing <THIS-CONFIG> with the config you are enabling or disabling and <SUBCOMMAND> with the command this config applies to.

Boolean config structure:

$ dbt --<THIS-CONFIG> <SUBCOMMAND>
$ dbt --no-<THIS-CONFIG> <SUBCOMMAND>

Boolean config example:

$ dbt --version-check run
$ dbt --no-version-check run

Environment variables contain a DBT_ prefix

$ export DBT_<THIS-CONFIG>=True
$ dbt run

You can set profile (or user) configurations in the config: block of profiles.yml. You would use the profile config to set defaults for all projects running on your local machine.

config:
  <THIS-CONFIG>: true

:::caution Experimental config flag This should be considered an unstable config flag because it is experimental and subject to change. We reserve the right to make breaking changes to this config flag. :::

At the start of runs, dbt caches metadata about all the objects in all the schemas where it might materialize resources (such as models). By default, dbt caches all schemas related to the project. When this config is enabled, dbt will only cache schemas related to selected resources for the current run. This can offer significant speed improvements when running a small subset of a large project.

For example, to improve speed and performance while focused on developing Salesforce models, which are materialized into their own dedicated schema, you would select those models and pass the cache-selected-only flag:

$ dbt --cache-selected-only run --select salesforce

The previous example enables you to start working with your salesforce models, and dbt will only cache those models instead of the entire project.

To set the config in your profile.yml:

config:
  cache_selected_only: true

Projects are recommended to set dbt version requirements, especially if they use features that are newer, or which may break in future versions of dbt Core. By default, if you run a project with an incompatible dbt version, dbt will raise an error.

You can use the VERSION_CHECK config to disable this check and suppress the error message:

$ dbt --no-version-check run
Running with dbt=1.0.0
Found 13 models, 2 tests, 1 archives, 0 analyses, 204 macros, 2 operations....

The DEBUG config redirects dbt's debug logs to standard out. The has the effect of showing debug-level log information in the terminal in addition to the logs/dbt.log file. This output is verbose.

The --debug flag is also available via shorthand as -d.

$ dbt --debug run
...

With the USE_EXPERIMENTAL_PARSER config, you can opt into the latest and greatest experimental version of the static parser, which is still being sampled for 100% correctness. See the docs on parsing for more details.

config:
  use_experimental_parser: true

Supply the -x or --fail-fast flag to dbt run to make dbt exit immediately if a single resource fails to build. If other models are in-progress when the first model fails, then dbt will terminate the connections for these still-running models.

For example, you can select four models to run, but if a failure occurs in the first model, the failure will prevent other models from running:

$ dbt -x run --threads 1
Running with dbt=1.0.0
Found 4 models, 1 test, 1 snapshot, 2 analyses, 143 macros, 0 operations, 1 seed file, 0 sources

14:47:39 | Concurrency: 1 threads (target='dev')
14:47:39 |
14:47:39 | 1 of 4 START table model test_schema.model_1........... [RUN]
14:47:40 | 1 of 4 ERROR creating table model test_schema.model_1.. [ERROR in 0.06s]
14:47:40 | 2 of 4 START view model test_schema.model_2............ [RUN]
14:47:40 | CANCEL query model.debug.model_2....................... [CANCEL]
14:47:40 | 2 of 4 ERROR creating view model test_schema.model_2... [ERROR in 0.05s]

Database Error in model model_1 (models/model_1.sql)
  division by zero
  compiled SQL at target/run/debug/models/model_1.sql

Encountered an error:
FailFast Error in model model_1 (models/model_1.sql)
  Failing early due to test failure or runtime error

The LOG_FORMAT config specifies how dbt's logs should be formatted. If the value of this config is json, dbt will output fully structured logs in JSON format; otherwise, it will output text-formatted logs that are sparser for the CLI and more detailed in logs/dbt.log.

$ dbt --log-format json run
{"code": "A001", "data": {"v": "=1.0.0"}, "invocation_id": "1193e449-4b7a-4eb1-8e8e-047a8b3b7973", "level": "info", "log_version": 1, "msg": "Running with dbt=1.0.0", "node_info": {}, "pid": 35098, "thread_name": "MainThread", "ts": "2021-12-03T10:46:59.928217Z", "type": "log_line"}

:::tip Tip: verbose structured logs

Use json formatting value in conjunction with the DEBUG config to produce rich log information which can be piped into monitoring tools for analysis:

$ dbt --debug --log-format json run

See structured logging for more details.

:::

The PARTIAL_PARSE config can turn partial parsing on or off in your project. See the docs on parsing for more details.

config:
  partial_parse: true

dbt --no-partial-parse run

By default, dbt will print out lines padded to 80 characters wide. You can change this setting by adding the following to your profiles.yml file:

config:
  printer_width: 120

We want to build the best version of dbt possible, and a crucial part of that is understanding how users work with dbt. To this end, we've added some simple event tracking to dbt (using Snowplow). We do not track credentials, model contents or model names (we consider these private, and frankly none of our business).

Usage statistics are fired when dbt is invoked and when models are run. These events contain basic platform information (OS + python version). You can see all the event definitions in tracking.py.

By default this is turned on – you can opt out of event tracking at any time by adding the following to your profiles.yml file:

config:
  send_anonymous_usage_stats: False

You can also use the DO_NOT_TRACK environmental variable to enable or disable sending anonymous data. For more information, see Environmental variables.

DO_NOT_TRACK=1 is the same as DBT_SEND_ANONYMOUS_USAGE_STATS=False DO_NOT_TRACK=0 is the same as DBT_SEND_ANONYMOUS_USAGE_STATS=True

The STATIC_PARSER config can enable or disable use of the static parser. See the docs on parsing for more details.

config:
  static_parser: true

As of v1.0, the -S or --strict flag has been deprecated.

By default, dbt shows all logs in standard out (stdout). You can use the QUIET config to show only error logs in stdout. Logs will still include the output of anything passed to the print() macro. For example, you might suppress all but error logs to more easily find and debug a jinja error.

config:
  quiet: true

Supply the -q or --quiet flag to dbt run to show only error logs and suppress non-error logs.

$ dbt --quiet run
...

By default, dbt includes print() messages in standard out (stdout). You can use the NO_PRINT config to prevent these messages from showing up in stdout.

config:
  no_print: true

Supply --no-print flag to dbt run to suppress print() messages from showing in stdout.

$ dbt --no-print run
...

By default, dbt will colorize the output it prints in your terminal. You can turn this off by adding the following to your profiles.yml file:

config:
  use_colors: False

$ dbt --use-colors run
$ dbt --no-use-colors run

Turning on the WARN_ERROR config will convert dbt warnings into errors. Any time dbt would normally warn, it will instead raise an error. Examples include --select criteria that selects no resources, deprecations, configurations with no associated models, invalid test configurations, or tests and freshness checks that are configured to return warnings.

$ dbt --warn-error run
...

The WRITE_JSON config determines whether dbt writes JSON artifacts (eg. manifest.json, run_results.json) to the target/ directory. JSON serialization can be slow, and turning this flag off might make invocations of dbt faster. Alternatively, you might disable this config if you want to perform a dbt operation and avoid overwriting artifacts from a previous run step.

dbt --no-write-json run