Reference
Helm Chart Values
Command Lines
This section shows the command lines for all of the ModelOps servers.
Data Channel Registry
The data channel registry is a server that supports data channel discovery and provides an API for querying available data channels.
The data channel registry is shipped in
<product-home>/distrib/tibco/modelops/bin/data-channel.jar
These command line options are supported:
-d,--local-data <zero or more local data files> Operate the server outside a Kubernetes environment, with local data files, for API testing only
-h,--help Display usage and exit
-l,--logging <Logback configuration file path> Set a non-default Logback configuration file to use for logging
-n,--namespace <Kubernetes namespace name> Set a non-default Kubernetes namespace in which the server will operate. Default is
no namespace, option with no argument discovers the namespace.
-p,--listen-port <listen port number Set a non-default listen port. Default value is 8080.
-t,--request-response-trace <"disable" | "basic" | "headers"> Manage request/response tracing
-v,--version Display version and exit
The data channel registry can be started using:
java -jar data-channel-registry.jar
Metrics
The metrics server provides aggregated real-time access to metrics captured in Prometheus. It is a Streaming node hosting LiveView.
The metrics application
is shipped as
<product-home>/distrib/tibco/modelops/application/metric-application.zip
These substitution variables are supported.
Name | Type | Description | Default |
---|---|---|---|
METRIC_DATA_MAX_AGE_MINUTES |
String | Maximum metric age in minutes | 5 |
METRIC_DATA_MAX_TABLE_SIZE_MEGABYTES |
String | Maximum metric table size in megabytes | 50 |
METRIC_DATA_POLLING_SECOND |
String | Metric data polling interval in seconds | 30 |
METRIC_METADATA_POLLING_SECOND |
String | Meta-data polling interval in seconds | 60 |
METRIC_PROMETHEUS_URL |
String | Prometheus server URL | http://localhost:9090 |
METRIC_SHARED_MEMORY_TYPE |
Enumeration | Shared memory type - FILE or SYSTEM_V_SHARED_MEMORY |
FILE |
The real-time metrics server can be started with default configuration values using:
epadmin install node --application metric-application.zip --nodename A.metrics
epadmin --servicename A.metrics start node
epadmin --servicename A.metrics stop node
epadmin --servicename A.metrics remove node
ModelOps
The ModelOps server supports artifact management and provides REST API to access all ModelOps functions.
The ModelOps server is shipped in
<product-home>/distrib/tibco/modelops/bin/modelops-server.jar
These command line options are supported:
-b,--backup <directory> backup the database and exit (backup file place in the directory)
-c,--configuration <file> configuration file
-f,--force remove records prior to restore (valid with restore only)
-gs,--generate-secret extract the master secret from the keystore into a file and exit
-h,--help display usage and exit
-kf,--keystore-file <file> keystore file (valid and required for generate-secret only)
-kp,--keystore-password <pass-phrase> keystore password (valid and required for generate-secret only)
-l,--logging <file> change default logging configuration
-ms,--master-secret <file> master secret file (valid and required for generate-secret only)
-r,--restore <file> restores the database from the backup file and exit
-s,--substitutions <name=value,...> substitution values
-sf,--substitution-file <file> substitution file
-t,--request-response-trace enable request/response tracing
-v,--version display version and exit
The ModelOps server can be started using:
java -jar modelops-server.jar
Scheduling
The Scheduling server provides support for starting jobs to execute scoring pipelines and data channels.
The Scheduling server is shipped in
<product-home>/distrib/tibco/modelops/bin/scheduling-server.jar
These command line options are supported:
-a,--admin-username <admin user name> Set the admin user name, defaults to admin
-c,--cleanup-age <cleanup time in minutes> Set cleanup time of completed tasks, defaults to 60 minutes
-g,--log-level <Log level> Set log level for the deployed scheduling job, defaults to INFO
-h,--help Display usage and exit
-l,--logging <Logback configuration file path> Set a non-default Logback configuration file to use for logging
-n,--namespace <Kubernetes namespace name> Set a non-default Kubernetes namespace in which the server will operate, defaults to current
namespace
-o,--local Operate the server outside a Kubernetes environment, with no scheduling, for API testing only
-p,--listen-port <listen port number> Set a non-default listen port. Default value is 8080.
-s,--session-timeout <session timeout in seconds> Set login session timeout in seconds, defaults to 3600 (1 hour)
-t,--request-response-trace <"disable" | "basic" | "headers"> Manage request/response tracing
-ti,--tools-image <tools container image name> Set tools container image name to be used when scheduling jobs in the future
-v,--version Display version and exit
The Scheduling server can be started using:
java -jar scheduling-server.jar
Scoring
The Scoring server supports model execution via model runners.
The Scoring server is shipped in
<product-home>/distrib/tibco/modelops/bin/scoring-server.jar
These command line options are supported:
-c,--configuration <Server configuration file path> Set a file to configure the server
-h,--help Display usage and exit
-l,--logging <Logback configuration file path> Set a non-default Logback configuration file to use for logging
-p,--listen-port <listen port number> Set a non-default listen port. Default value is 8080.
-r,--runners <List of runner directories> Set a list of directories to search for scoring runners, default
value is /opt/work/sb10/distrib/tibco/modelops/runner
-s,--session-timeout <session timeout in seconds> Set login session timeout in seconds, defaults to 3600 (1 hour)
-sf,--substitution-file <a file containing substitution variable settings in Java properties format> Provide a file of substitution
name/value pairs for use in a server configuration file
-sv,--substitutions <substitutions as comma-separated name=value pairs> Provide substitution name/value pairs for use in a server
configuration file
-t,--request-response-trace <"disable" | "basic" | "headers"> Manage request/response tracing
-v,--version Display version and exit
The Scoring server can be started using:
java -jar scoring-server.jar