Skip to content
RelationalAI Docs

Manage the RAI Native App for Snowflake

Manage the app

Section titled “Manage the app”

Follow these guides to manage the RAI Native App after installation:

Start the appBring the app online and confirm it is ready.
Configure user accessControl who can use the app and with what permissions.
Set up Direct AccessGet the lowest-latency connection to the RAI Native App.
View the app's statusConfirm whether the app is ready and healthy.
View logsView app logs and share support logs when needed.
Enable app observabilityRegister a secure events view so the app can report usage metrics.
Upgrade the appUnderstand how upgrades work and how to manage them.
Shut down the appStop the app to reduce costs during idle periods.
Understand costsSee what drives app costs and where you can reduce them.
Resources the app createsInventory the Snowflake resources tied to the RAI Native App.

Manage data sources

Section titled “Manage data sources”

Use these guides to understand how data sources are shared with the RAI Native App and how to manage them:

Prepare data sourcesPrepare a Snowflake table or view for use with RelationalAI.
Enable the CDC serviceSet up the service used to keep PyRel models in sync with source data.
Manage data streamsInspect and manage the data streams that synchronize Snowflake source data with PyRel models.
Fix data stream issuesRecover a data stream that stopped syncing.

Manage reasoners

Section titled “Manage reasoners”

Reasoners are the core compute resource in the RAI Native App that process PyRel semantic models and evaluate queries. Use these guides to understand how reasoners work and how to manage them:

Monitor reasoner usageMonitor reasoner health and resource use.
Manage reasoner jobsInspect or stop work on a specific reasoner.
Manage logic reasonersLearn how to configure and manage logic reasoners.
Manage prescriptive reasonersLearn how to configure and manage prescriptive reasoners.
Manage predictive reasonersLearn how to configure and manage predictive reasoners.
Enable warm reasonersReduce cold-start latency for logic reasoners.
Scale reasoner capacityIncrease or decrease reasoner capacity to meet workload demands.