Environment Variables#

Jina uses a number of environment variables to determine different behaviours. To see all supported environment variables and their current values, run

jina -vf

If you use containerized Executors (including Kubernetes and Docker Compose), you can pass separate environment variables to each Executor in the following way:

Include env vars in YAML

jtype: Flow
version: '1'
with: {}
executors:
- name: executor0
  port: 49583
  env:
    JINA_LOG_LEVEL: DEBUG
    MYSECRET: ${{ ENV.MYSECRET }}
- name: executor1
  port: 62156
  env:
    JINA_LOG_LEVEL: INFO
    CUDA_VISIBLE_DEVICES: 1

Include env vars in Python

from jina import Flow
import os

secret = os.environ['MYSECRET']
f = (
    Flow()
    .add(env={'JINA_LOG_LEVEL': 'DEBUG', 'MYSECRET': secret})
    .add(env={'JINA_LOG_LEVEL': 'INFO', 'CUDA_VISIBLE_DEVICES': 1})
)
f.save_config("envflow.yml")

See Also

For more information about the environment variable syntax used in Jina YAML configurations, see here.

The following environment variables are used internally in Jina:

Environment variable	Description
`JINA_AUTH_TOKEN`	Authentication token of Jina Cloud
`JINA_DEFAULT_HOST`	The default host where the server is exposed
`JINA_DEFAULT_TIMEOUT_CTRL`	The default timeout time used by Flow to check the readiness of Executors
`JINA_DEPLOYMENT_NAME`	The name of the deployment, used by the Head Runtime in Kubernetes to connect to different deployments
`JINA_DISABLE_UVLOOP`	If set, Jina will not use uvloop event loop for concurrent execution
`JINA_FULL_CLI`	If set, all the CLI options will be shown in help
`JINA_GATEWAY_IMAGE`	Used when exporting a Flow to Kubernetes or docker-compose to override the default gateway image
`JINA_GRPC_RECV_BYTES`	Set by the grpc service to keep track of the received bytes
`JINA_GRPC_SEND_BYTES`	Set by the grpc service to keep track of the sent bytes
`JINA_LOG_CONFIG`	The configuration used for the logger
`JINA_LOG_LEVEL`	The logging level used: INFO, DEBUG, WARNING
`JINA_LOG_NO_COLOR`	If set, disables color from rich console
`JINA_MP_START_METHOD`	Sets the multiprocessing start method used by jina
`JINA_RANDOM_PORT_MAX`	The max port number used when selecting random ports to apply for Executors or gateway
`JINA_RANDOM_PORT_MIN`	The min port number used when selecting random ports to apply for Executors or gateway
`JINA_LOCKS_ROOT`	The root folder where file locks for concurrent Executor initialization
`JINA_OPTOUT_TELEMETRY`	If set, disables telemetry
`JINA_K8S_ACCESS_MODES`	Configures the access modes for the PersistentVolumeClaim attached to the StatefulSet, when creating a StatefulSet in Kubernetes for an Executor using volumes. Defaults to ‘[“ReadWriteOnce”]’
`JINA_K8S_STORAGE_CLASS_NAME`	Configures the storage class for the PersistentVolumeClaim attached to the StatefulSet, when creating a StatefulSet in Kubernetes for an Executor using volumes. Defaults to ‘standard’
`JINA_K8S_STORAGE_CAPACITY`	Configures the capacity for the PersistentVolumeClaim attached to the StatefulSet, when creating a StatefulSet in Kubernetes for an Executor using volumes. Defaults to ‘10G’
`JINA_STREAMER_ARGS`	Jina uses this variable to inject GatewayStreamer arguments into the host environment running a Gateway