String representing the environment - dev, prod, etc. Used in labelling of resources. The only special value is 'local'. If environment is specified as 'local', a NodePort is used instead of ClusterIP to expose services.
Domain name pointed to the cluster. Eg. 'lulc.example.com' - You will need to create an A record for a wildcard of this domain to point to the ClusterIP of your Kubernetes cluster. For eg. if your domain is lulc.example.com, you will need to create an A record for *.lulc.example.com to the Cluster IP. This will then automatically create domains for the various services: api.lulc.example.com, socket.lulc.example.com, tiles.lulc.example.com.
Email address for cluster admin. Currently used only as the email address for Lets Encrypt when issuing SSL certificates.
Whether the API service should be enabled or not. Can be useful for debugging.
Number of replicas of the API container to run. Defaults to 2
Port that the service runs on inside the container. Ideally you should not need to change this.
If using GPUs, the number of total GPU instances available on the cluster - used by the API to calculate available instances, etc.
If using CPUs, the number of total CPU instances available to do model training / inference.
Base URL for Auth0 account for authentication. Example: https://pearl-landcover-staging.us.auth0.com
Random secret used by the API to sign tokens. Also used by the socket server to verify tokens. Should be a secure random string.
Postgres URL to access Postgres instance. This would typically be the URL provided by your managed postgres instance. Should include username and password. Example: postgres://user:[email protected]:5432/lulcdb
You should not need to edit this, will use the image specified by the upstream chart. Change only if you wish to specify a custom image for this container.
Tag for docker image. Again, this is specified by upstream. Change this only if you wish to use a custom image / tag combination.
No of CPUs requested, eg. 4, or 4000m
RAM requested - eg. 4Gi
No of CPUs limit
Memory limit
Whether the socket server is enabled or not. Defaults to true. Change this only for debugging purposes.
Port that the service runs on inside the container. You should not normally need to change this.
You should not need to edit this, will use the image specified by the upstream chart. Change only if you wish to specify a custom image for this container.
Tag for docker image. Again, this is specified by upstream. Change this only if you wish to use a custom image / tag combination.
No of CPUs requested, eg. 4
RAM requested - eg. 4Gi
No of CPUs limit
Memory limit
Whether the TiTiler server is enabled. Again, only set to false for debugging.
Port that the service runs on inside the container. You should not need to change this.
No of concurrent requests each TiTiler container can serve. Defaults to 4.
Whether to run TiTiler in Debug mode. Not recommended for production, but will give more useful log outputs.
Backend to store tile mosaics. Default: sqlite:///
URL for database host for storing mosaics. Default: /tmp/data/mosaics.db
Whether to use a node selector for the Tiler instances to force them onto nodes matching given label
Key to use for node selector
Value to use for node selector
You should not need to edit this, will use the image specified by the upstream chart. Change only if you wish to specify a custom image for this container.
Tag for docker image. Again, this is specified by upstream. Change this only if you wish to use a custom image / tag combination.
No of CPUs requested, eg. 4
RAM requested - eg. 4Gi
No of CPUs limit
Memory limit
Enable Redis container, used by API to maintain a cache. Is required by the API, so set to false only for debugging purposes.
No of CPUs requested, eg. 4
RAM requested - eg. 4Gi
No of CPUs limit
Memory limit
You should not need to edit this, will use the image specified by the upstream chart. Change only if you wish to specify a custom image for this container.
Tag for docker image. Again, this is specified by upstream. Change this only if you wish to use a custom image / tag combination.
Whether to enable placeholder containers, which ensure faster startup time for inference containers. You can disable this if startup time of inference containers is not important for your use-case.
Number of placeholder replicas to run.
Name of node selector to use to ensure placeholders run on same nodes as inference containers.
Whether running on GPU instances
Number of GPUs per instance in your cluster. Used to define size of the placeholder container.
No of CPUs requested, eg. 4
RAM requested - eg. 4Gi
No of CPUs limit
Memory limit
Whether to enable an Nginx server to redirect requests to a www. subdomain. For eg. redirect requests from example.com to www.example.com, where www.example.com is configured separately to serve the frontend code. You should normally not need this and can leave disabled.
Frontend domain, to redirect to www. - eg: example.com
You should not need to edit this, will use the image specified by the upstream chart. Change only if you wish to specify a custom image for this container.
Tag for docker image. Again, this is specified by upstream. Change this only if you wish to use a custom image / tag combination.
Used to cleanup inference containers that have not been cleanly terminated by the API. See https://github.com/hjacobs/kube-janitor for full options.