Usage of /usr/bin/agent: --component string The component name (predictor, explainer, transformer) to add as header to log events --component-port int Component port (default 8080) --config-dir string directory for model config files (default "/mnt/configs") --enable-batcher Enable request batcher --enable-puller Enable model puller --endpoint string The endpoint name to add as header to log events --inference-service string The InferenceService name to add as header to log events --log-batch-interval duration Max time to wait before flushing a partial batch --log-batch-size int Number of log records per batch for blob storage (default 1) --log-marshaller-port int Port for the embedded log marshaller HTTP server (default 9083) --log-marshaller-url string URL of the log marshaller service (default "http://localhost:9083/marshal") --log-mode string Whether to log 'request', 'response' or 'all' (default "all") --log-store-format string Output format for the log marshaller (json, csv, parquet) (default "json") --log-store-path string The path to the log output --log-url string The URL to send request/response logs to --logger-ca-cert-file string The logger CA certificate file (default "service-ca.crt") --logger-tls-skip-verify Skip verification of TLS certificate --max-batchsize string Max Batch Size (default "32") --max-latency string Max Latency in milliseconds (default "5000") --metadata-annotations strings Allow list of metadata annotation to be passed with payload logging --metadata-headers strings Allow list of headers that will be passed down as metadata --model-dir string directory for model files (default "/mnt/models") --namespace string The namespace to add as header to log events --port string Agent port (default "9081") --probe-period duration run readiness probe with given timeout (default -1ns) --source-uri string The source URI to use when publishing cloudevents --workers int Number of workers (default 5)