env-defaults

# Default Harmony Environment Variables - do not edit this file locally.
# Override any of the values here by setting a value in a .env file at the
# root level of this repository. All of the default values in this file
# should work for developing locally on Mac OS X.

###########################################################################
#                             No defaults possible                        #
#                                                                         #
# Parameters must be overridden, no default value can be provided         #
###########################################################################

# Random string used to sign cookies that should be consistent between
# load-balanced instances of Harmony.  You can use
# `hexdump -n 128 -e '"%08x"' /dev/random`
# to generate a 128 byte random hex string (256 chars)
COOKIE_SECRET=

# Random string used for encrypting and decrypting Earthdata login tokens
# sent to backend services
# `hexdump -n 32 -e '"%08x"' /dev/random`
# to generate a 32 character random string
SHARED_SECRET_KEY=

# See the "OAuth 2 (Earthdata Login)" section below for a description
OAUTH_CLIENT_ID=
OAUTH_UID=
OAUTH_PASSWORD=

###########################################################################
#                             General Config                              #
#                                                                         #
# Settings to control how Harmony behaves                                 #
###########################################################################

# The host network interface to bind against. Set to 127.0.0.1 to limit to
# only on the localhost when running locally
HOST_BINDING='0.0.0.0'

# The port on which to run the Harmony frontend
PORT=3000

# The port on which to run the Harmony backend that receives service
# callbacks
BACKEND_PORT=3001

# The port on which a worker container listens for work from the manager container
WORKER_PORT=5001

# The base URL for the service callbacks. This needs to be accessible from
# wherever the service is run. The Harmony CI/CD scripts sets this to the
# backend load balancer when running on AWS. The default is set for
# running Harmony in Kubernetes.
CALLBACK_URL_ROOT=http://harmony:3001
# When running locally on a Mac or Linux (not in Kubernetes), use the appropriate value below.
# Mac OS X:
# CALLBACK_URL_ROOT=http://host.docker.internal:3001
# Linux:
# CALLBACK_URL_ROOT=http://localhost:3001

# The CMR Endpoint to use (e.g. URL for local, SIT, UAT, or production)
CMR_ENDPOINT=https://cmr.uat.earthdata.nasa.gov

# For testing: Whether to use Localstack instead of AWS S3.  Options are "true" or "false".
USE_LOCALSTACK=true

# Hostname for harmony to use to connect to Localstack
LOCALSTACK_HOST=localstack

# Hostname for K8S pods to connect to Localstack
LOCALSTACK_K8S_HOST=localstack

# If set to true, only include EDL cookies if the request is made via https.
USE_HTTPS=false

# Whether to run example service endpoints under /example.  Useful for
# testing HTTP backends without actual service calls.  See example/http-service.js
EXAMPLE_SERVICES=true

# Identifier so backends know which Harmony client submitted the request
CLIENT_ID=harmony-local

# When set to true log messages are logged as a text string instead of the default
# JSON format. Useful when running harmony locally and viewing logs via a terminal.
TEXT_LOGGER=true

# When set to true log all database queries performed by knex
DEBUG_KNEX=false

# Log level (error, warn, info, debug)
LOG_LEVEL=debug

# Default number of results in a page
DEFAULT_RESULT_PAGE_SIZE=2000

# Default number of jobs listed in a page
DEFAULT_JOB_LIST_PAGE_SIZE=10

# TODO change this to a smaller number when aggregating services are updated to handle paged catalogs
# See HARNONY-1178
AGGREGATE_STAC_CATALOG_MAX_PAGE_SIZE=1000000

# Maximum number of results in a page
MAX_PAGE_SIZE=2000

# Number of granules allowed for a synchronous request. When the request exceeds
# this number it will be processed asynchronously. If a service provides a
# value in services.yml for maximum_sync_granules it will override this value.
MAX_SYNCHRONOUS_GRANULES=1

# The maximum allowed granules in any request synchronous or asynchronous. If a service
# attempts to configure a value greater than this limit for either maximum_async_granules
# or maximum_async_granules we will override the configuration to limit to this value.
MAX_GRANULE_LIMIT=2100

# The threshold of the number of granules in a request that will trigger auto-pausing with preview
PREVIEW_THRESHOLD=100

# String to identify the type of environment.  Options:
#  - "development" for local development.
#  - "production" for deployments with a separate postgres server (SIT, UAT, Production).
#  - "test" for unit / integration tests
# This is a conventional identifier to switch application behavior but is
# currently only used to determine which database connection to use in
# db/knexfile.js and whether to deploy the alert manager
NODE_ENV=development

# Connection string to use to connect to the database.  Used for
# NODE_ENV="production" databases.  For development and test, this is
# ignored, using a sqlite3 file instead
DATABASE_URL=

# A bucket with brief lifecycle where temporary uploads (shapefiles) are stored while
# requests are in flight
UPLOAD_BUCKET=local-upload-bucket

# An IAM role that allows access to a bucket and prefix where outputs for asynchronous
# requests are stored by backend services.
SAME_REGION_ACCESS_ROLE=same-region-access-role

# The CMR concept ID for the group whose users can access the admin interface
ADMIN_GROUP_ID=AG1222736883-EEDTEST

# The time (in seconds) between invocations of the job reaper service
JOB_REAPER_PERIOD_SEC=360

# Jobs older than this many minutes that are running are checked by the job reaper
REAPABLE_JOB_AGE_MINUTES=60

# The time (in seconds) between invocations of the work reaper service
WORK_REAPER_PERIOD_SEC=360

# WorkItems and WorkflowSteps (in a terminal state) older than this many minutes are checked by the work reaper
REAPABLE_WORK_AGE_MINUTES=1440

# The time (in seconds) between invocations of the work failer service
WORK_FAILER_PERIOD_SEC=360

# WorkItems and associated job (in a running state) older than this many minutes are set to failed by the work failer
FAILABLE_WORK_AGE_MINUTES=240

# The number of seconds to allow a pod to continue processing an active request before terminating a pod
DEFAULT_POD_GRACE_PERIOD_SECS=14400

# The number of ms to wait between polling to check whether a synchronous request completed
SYNC_REQUEST_POLL_INTERVAL_MS=100

# Local development: Use the following to set the Kubernetes context used by start scripts
# minikube users should set it to "minikube"
KUBE_CONTEXT=docker-desktop

# How long (in ms) to let a worker run on a work item before timing out and reporting an error
# default is 4 hours
WORKER_TIMEOUT=14400000

###########################################################################
#                             Service Config                              #
#                                                                         #
# Variables defined in config/services.yml that get passed to backend     #
# services on invocation                                                  #
#                                                                         #
# IMAGE_PULL_POLICY configuration:                                        #
# In deployed environments the pull policy used for images is 'Always'    #
# to ensure an up to date image is pulled as part of deploying. When      #
# developing locally the pull policy used is 'IfNotPresent' to            #
# prevent Kubernetes from pulling the image and overwriting any locally   #
# built images. Note if you want to pull an image update rather than      #
# building it you need to run `docker pull <IMAGE>` to update your local  #
# copy.                                                                   #
###########################################################################

# The service runner image used as a sidecar for polling for work for
SERVICE_RUNNER_REQUESTS_CPU=128m
SERVICE_RUNNER_REQUESTS_MEMORY=128Mi
SERVICE_RUNNER_LIMITS_CPU=1024m
SERVICE_RUNNER_LIMITS_MEMORY=512Mi
SERVICE_RUNNER_IMAGE=harmonyservices/service-runner:latest

# backend services
HARMONY_GDAL_ADAPTER_IMAGE=ghcr.io/nasa/harmony-gdal-adapter:latest
HARMONY_GDAL_ADAPTER_REQUESTS_CPU=128m
HARMONY_GDAL_ADAPTER_REQUESTS_MEMORY=128Mi
HARMONY_GDAL_ADAPTER_LIMITS_CPU=128m
HARMONY_GDAL_ADAPTER_LIMITS_MEMORY=8Gi
HARMONY_GDAL_ADAPTER_INVOCATION_ARGS='python -m gdal_subsetter'

HARMONY_SERVICE_EXAMPLE_IMAGE=harmonyservices/service-example:latest
HARMONY_SERVICE_EXAMPLE_REQUESTS_CPU=128m
HARMONY_SERVICE_EXAMPLE_REQUESTS_MEMORY=128Mi
HARMONY_SERVICE_EXAMPLE_LIMITS_CPU=128m
HARMONY_SERVICE_EXAMPLE_LIMITS_MEMORY=512Mi
HARMONY_SERVICE_EXAMPLE_INVOCATION_ARGS='python -m harmony_service_example'

HARMONY_NETCDF_TO_ZARR_IMAGE=harmonyservices/netcdf-to-zarr:latest
HARMONY_NETCDF_TO_ZARR_REQUESTS_CPU=128m
HARMONY_NETCDF_TO_ZARR_REQUESTS_MEMORY=128Mi
HARMONY_NETCDF_TO_ZARR_LIMITS_CPU=128m
HARMONY_NETCDF_TO_ZARR_LIMITS_MEMORY=512Mi
HARMONY_NETCDF_TO_ZARR_INVOCATION_ARGS='python -m harmony_netcdf_to_zarr'

SWOT_REPROJECT_IMAGE=sds/swot-reproject:latest
SWOT_REPROJECT_REQUESTS_CPU=128m
SWOT_REPROJECT_REQUESTS_MEMORY=128Mi
SWOT_REPROJECT_LIMITS_CPU=128m
SWOT_REPROJECT_LIMITS_MEMORY=512Mi
SWOT_REPROJECT_INVOCATION_ARGS='python swotrepr.py'

VAR_SUBSETTER_IMAGE=sds/variable-subsetter:latest
VAR_SUBSETTER_REQUESTS_CPU=128m
VAR_SUBSETTER_REQUESTS_MEMORY=128Mi
VAR_SUBSETTER_LIMITS_CPU=128m
VAR_SUBSETTER_LIMITS_MEMORY=512Mi
VAR_SUBSETTER_INVOCATION_ARGS='python subsetter.py'

SDS_MASKFILL_IMAGE=sds/maskfill-harmony:latest
SDS_MASKFILL_REQUESTS_CPU=128m
SDS_MASKFILL_REQUESTS_MEMORY=128Mi
SDS_MASKFILL_LIMITS_CPU=128m
SDS_MASKFILL_LIMITS_MEMORY=512Mi
SDS_MASKFILL_INVOCATION_ARGS='python harmony_adapter.py'

TRAJECTORY_SUBSETTER_IMAGE=sds/trajectory-subsetter:latest
TRAJECTORY_SUBSETTER_REQUESTS_CPU=128m
TRAJECTORY_SUBSETTER_REQUESTS_MEMORY=128Mi
TRAJECTORY_SUBSETTER_LIMITS_CPU=128m
TRAJECTORY_SUBSETTER_LIMITS_MEMORY=512Mi
TRAJECTORY_SUBSETTER_INVOCATION_ARGS='python harmony_service/adapter.py'

PODAAC_CONCISE_IMAGE=ghcr.io/podaac/concise:sit
PODAAC_CONCISE_REQUESTS_CPU=128m
PODAAC_CONCISE_REQUESTS_MEMORY=128Mi
PODAAC_CONCISE_LIMITS_CPU=128m
PODAAC_CONCISE_LIMITS_MEMORY=512Mi
PODAAC_CONCISE_INVOCATION_ARGS='concise_harmony'

PODAAC_L2_SUBSETTER_IMAGE=ghcr.io/podaac/l2ss-py:sit
PODAAC_L2_SUBSETTER_REQUESTS_CPU=128m
PODAAC_L2_SUBSETTER_REQUESTS_MEMORY=128Mi
PODAAC_L2_SUBSETTER_LIMITS_CPU=128m
PODAAC_L2_SUBSETTER_LIMITS_MEMORY=512Mi
PODAAC_L2_SUBSETTER_INVOCATION_ARGS='./docker-entrypoint.sh'

PODAAC_PS3_IMAGE=podaac/podaac-cloud/podaac-shapefile-subsetter:latest
PODAAC_PS3_REQUESTS_CPU=128m
PODAAC_PS3_REQUESTS_MEMORY=128Mi
PODAAC_PS3_LIMITS_CPU=128m
PODAAC_PS3_LIMITS_MEMORY=512Mi

PODAAC_NETCDF_CONVERTER_IMAGE=podaac/podaac-cloud/podaac-netcdf-converter:latest
PODAAC_NETCDF_CONVERTER_REQUESTS_CPU=128m
PODAAC_NETCDF_CONVERTER_REQUESTS_MEMORY=128Mi
PODAAC_NETCDF_CONVERTER_LIMITS_CPU=128m
PODAAC_NETCDF_CONVERTER_LIMITS_MEMORY=512Mi

QUERY_CMR_IMAGE=harmonyservices/query-cmr:latest
QUERY_CMR_REQUESTS_CPU=128m
QUERY_CMR_REQUESTS_MEMORY=128Mi
QUERY_CMR_LIMITS_CPU=128m
QUERY_CMR_LIMITS_MEMORY=512Mi

GIOVANNI_ADAPTER_IMAGE=harmonyservices/giovanni-adapter:latest
GIOVANNI_ADAPTER_REQUESTS_CPU=128m
GIOVANNI_ADAPTER_REQUESTS_MEMORY=128Mi
GIOVANNI_ADAPTER_LIMITS_CPU=128m
GIOVANNI_ADAPTER_LIMITS_MEMORY=512Mi
GIOVANNI_ADAPTER_INVOCATION_ARGS='node tasks/giovanni-adapter/app/cli'

# The services to deploy locally. A comma-separated list of services that the bin/deploy-services
# script should attempt to deploy. By default only a couple of harmony example services are deployed.
# When specifying another service to be deployed make sure the name matches the prefix for the
# sidecar file in tasks/service-runner-config and the name of the service in that file. For example
# to deploy the SWOT reprojection service which has a sidecar file swot-reproject-sidecar.yaml
# and service-name of swot-reproject, specify swot-reproject. query-cmr is required for all services,
# so it will always be deployed, and does not need to be specified. Make sure if the image for the
# service is not publicly available that you have built the docker image locally, otherwise the
# service will fail to start.
LOCALLY_DEPLOYED_SERVICES=harmony-service-example,harmony-netcdf-to-zarr

# page size to use with CMR calls
CMR_MAX_PAGE_SIZE=2000

# Prefix before "harmonyservices/task-name" for built-in tasks like query-cmr, e.g. an ECR location
# If not blank, it should end in a slash if there is a slash before "harmony"
BUILT_IN_TASK_PREFIX=
# Version to use for all built-in tasks.  "latest" (local / sandbox), "sit", "uat", or "prod"
BUILT_IN_TASK_VERSION=latest

# AWS region where harmony is deployed, irrelevant for local development
AWS_DEFAULT_REGION=us-west-2

# The object store used for storing shapefiles. Only S3 is supported currently.
OBJECT_STORE_TYPE=s3

# The bucket where final service output data will be staged
# Default: A bucket created for use in localstack
STAGING_BUCKET=local-staging-bucket

# The bucket where intermediate service artifacts will be staged
# Default: A bucket created for use in localstack
ARTIFACT_BUCKET=local-artifact-bucket

# The DNS entry to use for the host exposing the Harmony back end API for callbacks. This
# default is appropriate for when Harmony is running in Kubernetes.
BACKEND_HOST=harmony
#  For local development outside of Kubernetes:
# Mac OS X:
# BACKEND_HOST=host.docker.internal
# Linux:
# BACKEND_HOST=localhost

# Maximum number of non-file fields to accept when providing a shapefile to harmony
MAX_POST_FIELDS=100

# Maximum size (in bytes) for shapefiles
MAX_POST_FILE_SIZE=2000000000

# Maximum number of multipart parts to accept when providing a shapefile
MAX_POST_FILE_PARTS=100

# The location on the file system to save the work items that will be mounted into the
# containers for any services to read and save STAC catalogs.
HOST_VOLUME_PATH=

###########################################################################
#                             Prometheus Config                           #
#                                                                         #
# This section is irrelevant for local development unless you are a       #
# harmony core developer who plans to deploy Prometheus locally           #
###########################################################################

# For Prometheus time variables, "m" postfix means "minutes".
# https://prometheus.io/docs/prometheus/latest/querying/basics/#time-durations
# For k8s CPU or memory variables, see this link for unit explanations:
# https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/#meaning-of-cpu

# Prometheus deployment variables
PROMETHEUS_REQUESTS_CPU=128m
PROMETHEUS_REQUESTS_MEMORY=150Mi
PROMETHEUS_REQUESTS_EPHEMERAL_STORAGE=600Mi
PROMETHEUS_LIMITS_CPU=128m

PROMETHEUS_LIMITS_MEMORY=600Mi
PROMETHEUS_LIMITS_EPHEMERAL_STORAGE=2000Mi
PROMETHEUS_PROMETHEUS_SCRAPE_INTERVAL=15s
PROMETHEUS_POD_MANAGER_SCRAPE_INTERVAL=15s

# See https://prometheus.io/docs/prometheus/latest/storage/ for details on retention metrics.
# How long to keep prometheus metrics before cleaning up
PROMETHEUS_RETENTION_TIME=180d
# Maximum amount of space to use for Prometheus metrics before deleting oldest metrics
# Be sure to set this value to less than PROMETHEUS_LIMITS_EPHEMERAL_STORAGE otherwise the
# pod may be evicted due to running out of ephemeral space.
PROMETHEUS_RETENTION_SIZE=1750MB

# how many consecutive minutes to sample in the PromQL query for KubernetesPodNotHealthy
PROMETHEUS_POD_NOT_HEALTHY_DURATION=10m

# how long the specified alert must be firing before being sent to the Alert Manager
PROMETHEUS_POD_NOT_HEALTHY_WAIT_FOR=5m
PROMETHEUS_NODE_NOT_READY_WAIT_FOR=10m

# Alert manager variables (irrelevant to local development).
# To see/test whether alerts are firing locally, kubectl port-forward {prometheus-pod} 9090:9090 -n monitoring
# and navigate to the prometheus UI http://localhost:9090/alerts
ALERT_MANAGER_REQUESTS_CPU=500m
ALERT_MANAGER_REQUESTS_MEMORY=500M
ALERT_MANAGER_LIMITS_CPU=1
ALERT_MANAGER_LIMITS_MEMORY=1Gi

ALERT_MANAGER_TOPIC_ARN=

###########################################################################
#             Horizontal Pod Autoscaling Config                           #
#                                                                         #
# Variables that are used to configure scaling for service pods           #
#                                                                         #
###########################################################################

HPA_MIN_REPLICAS=1
HPA_MAX_REPLICAS=10

# See https://kubernetes.io/docs/reference/kubernetes-api/common-definitions/quantity/#Quantity
# for an explanation of the "m" suffix (e.g. 1.5 will be serialized as "1500m")
HPA_TARGET_VALUE=10000m

#############################################################################
#                        OAuth 2 (Earthdata Login)                          #
#                                                                           #
# Variables used to configure OAuth 2 authentication for Harmony data users #
#                                                                           #
# To use Earthdata Login, you must first set up a new application using the #
# Earthdata Login UI.                                                       #
# https://wiki.earthdata.nasa.gov/display/EL/How+To+Register+An+Application #
#############################################################################

# Variable            Description                 Example
#--------------------------------------------------------------------------------------
# OAUTH_CLIENT_ID     EDL App Client ID           rCjHBluumLhppiIX5iZxoQ
# OAUTH_UID           EDL App UID                 harmony_kbeam
# OAUTH_PASSWORD      EDL App Password            top_secret
# OAUTH_REDIRECT_URI  Valid EDL App Redirect URL  http://localhost:3000/oauth2/redirect
# OAUTH_HOST          Earthdata Login URL         https://uat.urs.earthdata.nasa.gov

# Notes:
# The OAUTH_REDIRECT_URI needs to be provided verbatim when setting up the
# Earthdata Login app. Each EDL app has a list of valid Redirect URLs, and
# Harmony requires a URI endpoint of "/oauth2/redirect" for all environments.
# When running Harmony locally, for example, add a Redirect URL of
# `http://localhost:3000/oauth2/redirect`. Use HTTPS for non-local instances.
#
# The OAUTH_HOST is used by the Harmony API as well as Harmony services.
# This should be the same EDL environment as used by backend data sources.
# For example, if OAUTH_HOST points to UAT EDL, downloading data from a
# TEA endpoint that's integrated with PROD EDL will fail.

OAUTH_REDIRECT_URI=http://localhost:3000/oauth2/redirect
OAUTH_HOST=https://uat.urs.earthdata.nasa.gov

###########################################################################
#                        Fallback Authentication                          #
###########################################################################

# When using data that cannot accept Earthdata Login tokens from
# Harmony, either because of application group setup or because they
# use a different Earthdata Login endpoint (e.g. Data needs production
# while Harmony and CMR need UAT), these provide a
# minimally-privileged EDL user login for fetching remote data behind
# Earthdata Login. This user should exist in all environments being
# used to fetch data.

# CAUTION: This fallback authentication switch should normally be
# disabled.

# CAUTION: This EDL user should *not* have accepted any EDL
# application EULAs. Doing so can result in users getting data for
# which they have not personally accepted the EULA.

# CAUTION: Enabling this could result in skewed metrics since
# downloads will appear to be done by Harmony--not the actual
# end-user--if the fallback authentication is used (e.g., a backend
# data source doesn't yet support EDL federated tokens).

FALLBACK_AUTHN_ENABLED=false
EDL_USERNAME=
EDL_PASSWORD=

###########################################################################
#                               Test Suite                                #
#                                                                         #
# Settings for altering the behavior of the test suite that are not used  #
# elsewhere in the application.                                           #
###########################################################################

# How to handle fixtures for remote calls in the test suite:
#   record (default): Perform and record new calls.  Replay existing calls.
#   bloody: Perform and record all calls.  Never play back.
#   replay: Always play back.  Throw an error if a call is not recorded.
#   cheat: Play back cached calls.  Perform and do not cache any new calls
REPLAY=record

# True if server logs should be allowed to go to STDOUT.  When false
# (default), they are routed to logs/test.log and suppressed in STDOUT
LOG_STDOUT=false