Prompt Proto Service LSSTComCamSim Helm values reference¶
Helm values reference table for the prompt-proto-service-lsstcomcamsim
application.
Key |
Type |
Default |
Description |
---|---|---|---|
prompt-proto-service.additionalVolumeMounts |
list |
|
Kubernetes YAML configs for extra container volume(s). Any volumes required by other config options are automatically handled by the Helm chart. |
prompt-proto-service.affinity |
object |
|
Affinity rules for the prompt processing pods |
prompt-proto-service.alerts.server |
string |
|
Server address for the alert stream |
prompt-proto-service.alerts.topic |
string |
None, must be set |
Topic name where alerts will be sent |
prompt-proto-service.alerts.username |
string |
|
Username for sending alerts to the alert stream |
prompt-proto-service.apdb.config |
string |
None, must be set |
URL to a serialized APDB configuration, or the “label:” prefix followed by the indexed name of such a config. |
prompt-proto-service.cache.baseSize |
int |
|
The default number of datasets of each type to keep. The pipeline only needs one of most dataset types (one bias, one flat, etc.), so this is roughly the number of visits that fit in the cache. |
prompt-proto-service.cache.maxFilters |
int |
|
The maximum number of datasets of a given type the service might load if the filter is unknown. Should be greater than or equal to the number of filters that have e.g. flats or transmission curves. |
prompt-proto-service.cache.patchesPerImage |
int |
|
A factor by which to multiply |
prompt-proto-service.cache.refcatsPerImage |
int |
|
A factor by which to multiply |
prompt-proto-service.containerConcurrency |
int |
|
The number of Knative requests that can be handled simultaneously by one container |
prompt-proto-service.image.pullPolicy |
string |
|
Pull policy for the PP image |
prompt-proto-service.image.repository |
string |
|
Image to use in the PP deployment |
prompt-proto-service.image.tag |
string |
|
Overrides the image tag whose default is the chart appVersion. |
prompt-proto-service.imageNotifications.imageTimeout |
int |
|
Timeout to wait after expected script completion for raw image arrival (seconds). |
prompt-proto-service.imageNotifications.kafkaClusterAddress |
string |
None, must be set |
Hostname and port of the Kafka provider |
prompt-proto-service.imageNotifications.topic |
string |
None, must be set |
Topic where raw image arrival notifications appear |
prompt-proto-service.instrument.calibRepo |
string |
None, must be set |
URI to the shared repo used for calibrations, templates, and pipeline outputs. If |
string |
|
The “short” name of the instrument |
|
prompt-proto-service.instrument.pipelines.main |
string |
None, must be set |
Machine-readable string describing which pipeline(s) should be run for which visits. Notation is complex and still in flux; see the source code for examples. |
prompt-proto-service.instrument.pipelines.preprocessing |
string |
None, must be set |
Machine-readable string describing which pipeline(s) should be run before which visits’ raw arrival. |
prompt-proto-service.instrument.preloadPadding |
int |
|
Number of arcseconds to pad the spatial region in preloading. |
prompt-proto-service.instrument.skymap |
string |
|
Skymap to use with the instrument |
prompt-proto-service.knative.cpuLimit |
int |
|
The maximum cpu cores for the full pod (see |
prompt-proto-service.knative.cpuRequest |
int |
|
The cpu cores requested for the full pod (see |
prompt-proto-service.knative.ephemeralStorageLimit |
string |
|
The maximum storage space allowed for each container (mostly local Butler). This allocation is for the full pod (see |
prompt-proto-service.knative.ephemeralStorageRequest |
string |
|
The storage space reserved for each container (mostly local Butler). This allocation is for the full pod (see |
prompt-proto-service.knative.extraTimeout |
int |
|
To acommodate scheduling problems, Knative waits for a request for twice |
prompt-proto-service.knative.gpu |
bool |
|
GPUs enabled. |
prompt-proto-service.knative.gpuRequest |
int |
|
The number of GPUs to request for the full pod (see |
prompt-proto-service.knative.idleTimeout |
int |
|
Maximum time that a container can send nothing to Knative (seconds). This is only useful if the container runs async workers. If 0, idle timeout is ignored. |
prompt-proto-service.knative.memoryLimit |
string |
|
The maximum memory limit for the full pod (see |
prompt-proto-service.knative.memoryRequest |
string |
|
The minimum memory to request for the full pod (see |
prompt-proto-service.knative.responseStartTimeout |
int |
|
Maximum time that a container can send nothing to Knative after initial submission (seconds). This is only useful if the container runs async workers. If 0, idle timeout is ignored. |
prompt-proto-service.logLevel |
string |
log prompt_processing at DEBUG, other LSST code at INFO, and third-party code at WARNING. |
Requested logging levels in the format of Middleware’s –log-level argument. |
prompt-proto-service.podAnnotations |
object |
See the |
Annotations for the prompt-proto-service pod |
prompt-proto-service.raw_microservice |
string |
|
The URI to a microservice that maps image metadata to a file location. If empty, Prompt Processing does not use a microservice. |
prompt-proto-service.registry.centralRepoFile |
bool |
|
If set, this application’s Vault secret must contain a |
prompt-proto-service.s3.auth_env |
bool |
|
If set, get S3 credentials from this application’s Vault secret. |
prompt-proto-service.s3.disableBucketValidation |
int |
|
Set this to disable validation of S3 bucket names, allowing Ceph multi-tenant colon-separated names to be used. |
prompt-proto-service.s3.endpointUrl |
string |
|
|
prompt-proto-service.s3.imageBucket |
string |
None, must be set |
Bucket containing the incoming raw images |
prompt-proto-service.sasquatch.auth_env |
bool |
|
If set, this application’s Vault secret must contain a |
prompt-proto-service.sasquatch.endpointUrl |
string |
|
Url of the Sasquatch proxy server to upload metrics to. Leave blank to disable upload. This is a preliminary implementation of Sasquatch support, and this parameter may be deprecated if we instead support |
prompt-proto-service.sasquatch.namespace |
string |
|
Namespace in the Sasquatch system with which to associate metrics. |
prompt-proto-service.worker.grace_period |
int |
|
When Knative shuts down a pod, the time its workers have to abort processing and save intermediate results (seconds). |
prompt-proto-service.worker.restart |
int |
|
The number of requests to process before rebooting a worker. If 0, workers process requests indefinitely. |
prompt-proto-service.worker.timeout |
int |
|
Maximum time that a worker can process a next_visit request (seconds). |