Carvel Logo

Data Values

Overview

A ytt run can be configured by supplying custom Data Values.

(For a high-level overview of ytt, see How it works.)

Declaring Data Values

Typically, Data Values are declared in a schema file. See the Using Data Values guide for more details.

Note: ytt continues to support declaring Data Values without schema for backwards-compatibility. However, due to the significantly improved support for catching configuration errors that schema brings, it is the recommended method for doing so.

Configuring Data Values

Data Values can be configured in one of two ways:

Configuring Data Values via command line flags

The --data-value... family of command-line flags provides a means of configuring Data Values from:

  • the command-line, itself;
  • OS environment variables;
  • “data values file"s — plain YAML files containing values for multiple Data Values

Those flags are:

--data-value [@lib:]key=value — sets a Data Value to a string value

  • key — name of Data Value. Use dot notation for nested values (e.g. key2.nested=val)
  • value — value to set (always interpreted as a string)
  • @lib: — (optional) specify library whose data values to configure (details below)
  • examples: instance.count=123, key=string, input=true, all set to strings

--data-value-yaml [@lib:]key=value) — sets a Data Value to a YAML-parsed value

  • key — name of Data Value.
  • value — value to set (decoded as a YAML value)
  • @lib: — (optional) specify library whose data values to configure (details below)
  • examples: instance.count=123 sets as integer, key=string as string, input=true as bool

--data-value-file [@lib:]key=file-path — sets a single Data Value to the contents of a given file.

  • key — name of Data Value.
  • file-path — file-system path to a file whose contents will become the value of key.
  • @lib: — (optional) specify library whose data values to configure (details below)
  • particularly useful for loading multi-line string values from files such as private and public key files, certificates, etc.
  • not to be confused with --data-values-file (described below).

--data-values-env [@lib:]PREFIX — sets one or more Data Values to string values from OS environment variables that start with the given prefix.

  • PREFIX — the literal prefix used to select the set of environment variables from which to configure Data Values.
  • @lib: — (optional) specify library whose data values to configure (details below)
  • for nested values, use double-underscore (i.e. __) in environment variable names to denote a “dot”.
  • example:
    with environment variables…
    DVAL_key1=blue
    DVAL_key2__nested=1337
    
    … and the parameter …
    $ ytt ... --data-values-env DVAL ...
    
    … would set two Data Values:
    key1=blue and
    key2.nested="1337". (both as strings)

--data-values-env-yaml [@lib:]PREFIX — sets one or more Data Values to the YAML-parsed values from OS environment variables that start with the given prefix.

  • PREFIX — the literal prefix used to select the set of environment variables from which to configure Data Values.
  • @lib: — (optional) specify library whose data values to configure (details below)
  • for nested values, use double-underscore (i.e. __) in environment variable names to denote a “dot”.
  • example:
    with environment variables…
    DVAL_key1=blue
    DVAL_key2__nested=1337
    
    … and the parameter …
    $ ytt ... --data-values-env DVAL ...
    
    … would set two Data Values:
    key1=blue (a string value) and
    key2.nested=1337 (an integer value).

--data-values-file [@lib:]file-path — sets one or more Data Values from a plain YAML file.

  • file-path — file-system path to a file which will be parsed as YAML structured identically to expected Data Values.
    • file must be plain YAML (i.e. not a ytt template or Data Values Overlay); it cannot contain YAML comments starting with #@.
    • array values replace (rather than append to) any existing value.
    • if there are more than one YAML documents in such a file, they are merged from top to bottom (last wins)
  • @lib: — (optional) specify library whose data values to configure (details below)
  • example:
    with the file prod-values.yml
    domain: example.com
    client_opts:
      timeout: 10
      retry: 5
    
    … and the parameter …
    $ ytt ... --data-values-file prod-values.yml ...
    
    … would set all three Data Values:
    domain=example.com (a string value)
    client_opts.timeout=10 (an integer value)
    client_opts.retry=5 (an integer value).

Notes:

  • As of v0.34.0+ Data Values passed via --data-value... flags do not necessarily need to be declared beforehand. In prior versions of ytt, a Data Value must be declared (back then, specified in a @data/values overlay, typically), before it could be configured through a flag.
  • the --data-value... flags can be repeated multiple times and used in any combination. See Data Values merge order for details on how they combine.
    export STR_VALS_key6=true # will be string 'true'
    export YAML_VALS_key6=true # will be boolean true
    
    ytt -f . \
      --data-value key1=val1-arg \
      --data-value-yaml key2.nested=123 \ # will be int 123
      --data-value-yaml 'key3.other={"nested": true}' \
      --data-value-file key4=/path \
      --data-values-env STR_VALS \
      --data-values-env-yaml YAML_VALS
    

Configuring Data Values via Data Values Overlays

Data Values can also be configured via a specific kind of ytt Overlay.

A Data Values Overlay is a YAML document annotated with @data/values.

#@data/values
---
key1: val1
key2:
  nested: val2
key3: val3
key4:

Note:

Splitting Data Values Overlays into multiple files

Available in v0.13.0.

It’s possible to split data values into multiple files (or specify multiple data values in the same file). @ytt:data library provides access to the merged result. Merging is controlled via overlay annotations and follows same ordering as overlays. Example:

values-default.yml:

#@data/values
---
key1: val1
key2:
  nested: val2
key3:
key4:

values-production.yml:

#@data/values
---
key3: new-val3
#@overlay/remove
key4:
#@overlay/match missing_ok=True
key5: new-val5

Note that key4 is being removed, and key5 is marked as missing_ok=True because it doesn’t exist in values-default.yml (this is a safety feature to prevent accidental typos in keys).

config.yml:

#@ load("@ytt:data", "data")

first: #@ data.values.key1
third: #@ data.values.key3
fifth: #@ data.values.key5

Running ytt -f . (or ytt -f config.yml -f values-default.yml -f values-production.yml) results in:

first: val1
third: new-val3
fifth: new-val5

See Multiple data values example in the online playground.

Data Values merge order

Data values are merged in following order (latter one wins):

  1. default values from @data/values-schema files
  2. @data/values overlays (same ordering as overlays)
  3. --data-values-file specified files (left to right)
  4. --data-values-env specified values (left to right)
  5. --data-values-env-yaml specified values (left to right)
  6. --data-value specified value (left to right)
  7. --data-value-yaml specified value (left to right)
  8. --data-value-file specified value (left to right)

(When configuring libraries, the data values merge order, is the same, even if through different mechanisms.)


Library data values

Available in v0.28.0+

Each library may specify data values which will be evaluated separately from the root level library.

Setting library values via files

To override library data values, add @library/ref annotation to data values YAML document, like so:

#@library/ref "@lib1"
#@data/values
---
key1: val1
key2: val2

#@library/ref "@lib1"
#@data/values after_library_module=True
---
key3: val3

The @data/values annotation also supports a keyword argument after_library_module. If this keyword argument is specified, given data values will take precedence over data values passed to the .with_data_values(...) function when evaluating via the library module.

Setting library values via command line flags

Data value flags support attaching values to libraries for use during library module evaluation:

export STR_VALS_key6=true # will be string 'true'
export YAML_VALS_key6=true # will be boolean true

ytt -f . \
  --data-value @lib1:key1=val1-arg \
  --data-value-yaml @lib2:key2.nested=123 \ # will be int 123
  --data-value-yaml '@lib3:key3.other={"nested": true}' \
  --data-value-file @lib4:key4=/path \
  --data-values-env @lib5:STR_VALS \
  --data-values-env-yaml @lib6:YAML_VALS
  --data-values-file @lib6:/path

Library Data Values merge order

For a given library instance, data values are merged in following order (latter one wins):

  1. default values from schema:
    1. @data/values-schema files within the library
    2. @data/values-schema files externally referenced in.
    3. given through instance.with_data_values_schema()
    4. @data/values-schema files externally referenced in after_library_module=True.
  2. values from data value sources:
    1. @data/values overlays within the library (same ordering as overlays)
    2. @data/values overlays externally referenced in
    3. specified using instance.with_data_values()
    4. @data/values overlays externally referenced in after_library_module=True
    5. --data-values-file specified files referenced in (left to right)
    6. --data-values-env specified values referenced in (left to right)
    7. --data-values-env-yaml specified values referenced in (left to right)
    8. --data-value specified value referenced in (left to right)
    9. --data-value-yaml specified value referenced in (left to right)

(Help improve our docs: edit this page on GitHub)