Due to Kubernetes ConfigMap
storage limitations, the total size of a pipeline and its associated resources
can’t exceed 1 MiB. This limit includes the combined size of configuration files,
secrets, parsers,
and any other files.
Modify a pipeline’s configuration file
Use one of these methods to modify a pipeline’s configuration file.- Web interface
- Pipeline CLI
When you modify a pipeline by using the Telemetry Pipeline web interface, that
pipeline’s configuration file updates to reflect your changes, even if you didn’t
edit the file directly.You can also edit configuration files directly in the Advanced Settings
section of the pipeline builder, either by pasting raw YAML into the text field
or by using Import to select a YAML file from your device. If the configuration
settings you specify are valid, the pipeline builder continuously saves any changes
you make and updates your pipeline accordingly.
Format
Configuration files follow standard YAML conventions and use the.yaml file
extension. Regardless of a pipeline’s individual settings, all configuration
files adhere to the same core format:
Configuration files use two spaces per indent, not four.
Sections
All configuration files include apipeline section, an inputs section, and
an outputs section. Additionally, each input or output can contain an associated
processors section.
Pipeline
Thepipeline section is the top-level section in a configuration file. It’s a
parent to the inputs section and the outputs section.
Inputs
Theinputs section is a sequence of mappings that contains one or more sources
of telemetry data. Each mapping corresponds to a single
source plugin, and includes
key/value pairs to configure the settings for that plugin, plus an optional
processors object.
Outputs
Theoutputs section is a sequence of mappings that contains one or more destinations
for your telemetry data. Each mapping corresponds to a single
destination plugin, and includes
key/value pairs to configure the settings for that plugin, plus an optional
processors object.
Processors
Eachprocessors section is a child to one of the mappings within the inputs
section or outputs section. These sections contain
processing rules to modify the
telemetry data passing through each input or output.
Because processing rules run separately for different types of telemetry data,
a processors section is grouped into subsections for logs, metrics, and
traces, where applicable. Each telemetry type subsection includes an
actions object, which is a sequence of mappings that contains one or more
processing rules and each rule’s associated settings.
If an input or output doesn’t have processing rules associated with a particular
type of telemetry data, the subsection for that telemetry type is omitted.