OLTP Classification Agent

The OLTP Classification Agent is a container ALTR provides that runs classification against your OLTP databases – Oracle, SQL Server, PostgreSQL, and MySQL – from inside your own environment. The agent connects to your database, samples data locally, evaluates your classifiers, and uploads only the results to ALTR.

Note

OLTP classification is enabled per account. Contact ALTR to enable OLTP classification and obtain the agent’s container image.

What the Agent Runs

The agent runs the same collections and classifiers as ALTR Hosted and In-Warehouse classification – build a collection once and it works wherever classification runs. See Classification Processing Location for which classifier rules evaluate on OLTP jobs.

OLTP data sources always classify through this agent – unlike Snowflake, which also offers ALTR Hosted and In-Warehouse processing, there’s no alternative processing location for OLTP.

File Sources

The agent also classifies files in an Amazon S3 bucket or a file system directory, registered as repositories alongside your databases. CSV, JSON, and XML files are read one field per header, key, or element, plain-text files line by line, and every collection the agent runs against a database runs against files unchanged. See File Classification for the formats ALTR reads and what a file report contains, and File Classification with Agents for setup.

Data Handling for External Conditions

Sampled data stays in your environment, with one exception: a classifier’s Google DLP or Amazon Comprehend Conditions condition calls the corresponding cloud service directly, using credentials you provide. Those calls go to your own Google Cloud or AWS account – never to ALTR.

Scheduling

The agent supports cron scheduling, so classification jobs can run on a recurring basis without manual intervention. Snowflake and Databricks classification jobs must be triggered on demand.

Installation and Registration

Setting up the agent – registering your database as a repository (ALTR’s term for a registered data source), generating a key pair, running the container, and configuring credential providers – is covered separately: see Data Classification with Agents.