Hybrid Runner environment variables reference

There are additional settings that can be configured through the use of these additional docker run environment variables.

Environment variable	Value
BODY_LIMIT	Define the maximum request body size. The default value is 30mb.
CUSTOM_API_ROUTE	Define additional path location to the resources in hybrid runner. Example: If this value is set to mypath/torun then the call to v3/execute would change from http://localhost:3000/mytenant/api/v3/execute to http://localhost:3000/mypath/torun/mytenant/api/v3/execute.
PORT	Set the service listening port. The default value is 3000.
UPLOAD_ENABLE	Allow Wasms to be uploaded to the Hybrid Runner directly using the /upload endpoint. This is not recommended to be used for production. The default value is false.

Manage access to filesystem

By default, the runner requires read-write permissions to the following paths.

Directory and path	Description
Tenant resolver /project/src/app/src/tenant_resolver	Purpose: Stores dynamically generated tenant-specific resolver JavaScript files. Write pattern: Created once per tenant on the first request, rarely modified thereafter. Example: externalResolver_<tenant_name>.js
Service map cache /project/src/app/src/services/servicemap From nodegen-server 1.50.1 and onwards /project/src/app/src/servicemap	Purpose: Caches service metadata (servicemap.json) to resolve models. This avoids repeated directory scanning. Write pattern: Generated on application startup in offline mode. Recreated when model location changes. Files created: servicemap.json
Temporary file storage /tmp	Purpose: Store and caches Wasm. This is used by Node.js as the tmp library working directory. Write pattern: Frequent writes during model downloads, cleaned up by graceful cleanup handlers. Cleanup: Periodic cleanup of files older than 7 days.

In nodegen-server versions >=1.50.2, the locations of these paths can be customized using environment variables. If the account does not have read-write permissions to the locations above, this can be used to adjust the default paths.

Environment variable	Value
CACHE_DIR	Cache write directory for Tenant resolver and Service map cache. Default value are the paths noted above. To set an absolute path, start the path using /, e.g. /var/runner_cache. To set a relative path within the working directory, start the path using an eligible folder character, e.g. runner_cache would be set to ./runner_cache.
TEMP_DIR	Temporary file storage directory. Default value is the path noted above. To set an absolute path, start the path using /, e.g. /tmp. To set a relative path within the working directory, start the path using an eligible folder character, e.g. tmp would be set to ./tmp.

Control log output

In nodegen-server versions >=1.50.2, more variables are available to manage log output.

Environment variable	Value
LOG_PATH	Default: Logs are recorded to stdout only. When defined, logs are saved to the defined location. To set an absolute path, start the path using /, e.g. /var/runner_logs. To set a relative path within the working directory, start the path using an eligible folder character, e.g. runner_logs would be set to ./runner_logs.
LOG_ROTATION_ENABLED	Default: true This removes files that exceed the maximum log storage criteria.
LOG_MAX_SIZE	Default: 10M Maximum file size before rotation: K for kilobytes. M for megabytes. G for gigabytes.
LOG_ROTATION_INTERVAL	Default: 1d Time-based rotation interval: h for hours. d for days. M for months.
LOG_MAX_FILES	Default: 14 Maximum number of rotated log files to keep.
LOG_COMPRESSION	Default: false Whether gzip compression is used on the log files.

Configure OpenTelemetry

Starting from nodegen-server versions >=1.46.0, the image implements support for OpenTelemetry. OpenTelemetry is an observability framework and toolkit to facilitate the generation, export, collection of telemetry data. The package conducts auto-instrumentation using @opentelemetry/auto-instrumentations-node. The data collected includes:

Traces	Metrics	Attributes
HTTP request/response details Route information Request duration Error details Correlation IDs Tenant information	Request counts Error rates Request duration Memory usage (heap, RSS)	Service name and version Environment HTTP method and route Status codes Request IDs

Below are links to documentation on integrating OpenTelemetry with popular Application Performance Monitoring (APM) applications:

• Amazon CloudWatch: OpenTelemetry

• Azure App Insights: Application Insights with OpenTelemetry

• Dynatrace: OpenTelemetry & Dynatrace for intelligent observability and OpenTelemetry and Dynatrace

• Elastic: What is OpenTelemetry? and Use OpenTelemetry with APM

• Sentry: Code-level application monitoring for OpenTelemetry and OpenTelemetry Support

The configuration of OpenTelemetry in the hybrid runner is managed through environment variables defined in the Docker container. Information on the relevant environment variables can be found in:

• SDK Configuration

• Environment Variable Specification

Key configuration variables include:

Environment variable	Value
OTEL_SDK_DISABLED	Disable the SDK for all signals.
OTEL_SERVICE_NAME	Sets the value of the service.name resource attribute.
OTEL_RESOURCE_ATTRIBUTES	Key-value pairs to be used as resource attributes.
OTEL_EXPORTER_OTLP_ENDPOINT	If using the OpenTelemetry Protocol Exporter: Target to which the exporter is going to send spans, metrics, or logs.
OTEL_EXPORTER_OTLP_HEADERS	If using the OpenTelemetry Protocol Exporter: Key-value pairs to be used as headers associated with gRPC or HTTP requests.

Manage compatible Neuron versions

The compatibility feature in the runner is included in nodegen-server >=1.3.0.

Neuron compiler versions >=1.22.1 embed the compiler version into the ZIP package. If this feature is enabled and compiler version cannot be found, then the runner assume the Neuron version will be compatible.

The list of Neuron compatibility versions that can be found in the repository nodegen-server-compatibility.

nodegen-server also has the ability to check whether Neuron Wasms are supported by the version of the runner you are using. This is done by downloading the JSON file from nodegen-server-compatibility and assessing it against the versions of the runner and the requested Wasm. This can only be done with an online connection to fetch the necessary compatibility files.

This can be configured through the use of these additional docker run environment variables.

Environment variable	Value
NODEGEN_COMPATIBILITY_BASE_URL	Base URL for Github user content. https://raw.githubusercontent.com
NODEGEN_COMPATIBILITY_PATH	Path to the JSON document in nodegen-server-compatibility. /Coherent-Partners/nodegen-server-compatibility/main/nodegen-server-compatibility.json
CHECK_RUNNER_COMPATIBILITY	true or false Whether to check if the version of the runner supports this version of Neuron. The default value is false.
EXECUTE_WITH_WARNING	true or false If CHECK_RUNNER_COMPATIBILITY is true, this defines whether or not to continue with executing the Spark service but with a warning. The default value is false.
NODEGEN_COMPATIBILITY_LIST_REFRESH_INTERVAL	If CHECK_RUNNER_COMPATIBILITY is true, this defines how often in minutes to fetch the nodegen-server-compatibility JSON document. The default value is 1440.