Protegrity Agent on

Prerequisites

Mon, 01 Jan 0001 00:00:00 +0000

The following requirements are met before installing Protegrity Agent with PPC.

The jumpbox is registered and prepared.
For more information about registering a jumpbox, refer to Configuring Authentication for Protegrity AI Team Edition.
Ensure that a PPC cluster is installed and accessible.
For more information about installing a PPC, refer Installing PPC.
Ensure that the Protegrity Policy Manager is installed. Install the Policy Workbench to deploy the Protegrity Policy Manager.

For more information about installing the Policy Workbench, refer to the section Installing Protegrity Policy Manager.

Configuring Protegrity Agent

Mon, 01 Jan 0001 00:00:00 +0000

API Service Endpoints

This section provides an overview of the service API endpoints exposed by Protegrity Agent.

Name	Endpoint
Conversations	`/pty/proagent/v1.0/conversations`
Responses	`/pty/proagent/v1.0/responses`
Health Check	`/pty/proagent/v1.0/health`
Readiness Probe	`/pty/proagent/v1.0/ready`
Liveness Probe	`/pty/proagent/v1.0/live`
Version Info	`/pty/proagent/v1.0/version`
UI Dashboard	`/pty/proagent/v1.0/ui`

Protegrity Agent Configurations

In addition to the OpenAI endpoints, the following parameters are configurable through a user-provided my-values.yaml file, supplied during deployment using the helm upgrade ... -f my-values.yaml command.

Section	Variable	Comment
global	`restore`	Restore mode flag. When `true`, Velero restores the backup PVC from S3 and a restore Job imports the `pg_dump` into a fresh Postgres instance.
`proagentService.environment`	`LOG_LEVEL`	Application log level (default: `INFO`)
`proagentService.environment`	`THINKING_TIMEOUT`	Agent pauses and asks for feedback if it goes over this limit (in minutes). Must be less than the application’s internal response timeout or any other network timeouts.
`backup`	`enabled`	Backup status
`backup`	`schedule`	`pg_dump` CronJob schedule
`backup`	`veleroSchedule`	Velero snapshot schedule (30-min offset)
`backup`	`scheduleName`	Backup name
`backup`	`backupName`	Set to a specific backup name for point-in-time restore

Updating the Deployment

To update the deployed cluster it is recommended to create a my-values.yaml and then helm upgrade ... -f my-values.yaml.

Uninstalling Protegrity Agent

Mon, 01 Jan 0001 00:00:00 +0000

To uninstall the Protegrity Agent, run the following command.

helm uninstall protegrity-agent -n pty-protegrity-agent
kubectl patch ec2nodeclass protegrity-agent-nodeclass -p '{"metadata":{"finalizers":[]}}' --type=merge
kubectl delete namespace pty-protegrity-agent

To verify the uninstallation is successfully completed, run the following command:

kubectl get all -n pty-protegrity-agent 2>/dev/null; kubectl get ec2nodeclass,nodepool,nodes

If the command returns no output, all tracked resources for the product are successfully removed from the cluster.

This operation might require few minutes for clearing all the resources.

Appendix - Features and Capabilities and Limitations

Mon, 01 Jan 0001 00:00:00 +0000

ProAgent Singleton

Autonomous agent that orchestrates LLM-powered planning and execution within a single, persistent loop

Maintains a per-conversation agent instance managed by an agent pool, ensuring full state isolation between concurrent conversations
Runs a manual iteration loop with dynamic context re-injection: reads workspace knowledge, plan state, and domain-specific skills before every LLM call
Routes queries via LLM classification between direct answers and structured plan-execute workflows
Coordinates the planner and executor to decompose complex requests into atomic steps and execute them sequentially
Dispatches tool calls for API interactions, file operations, knowledge updates, and human feedback requests
Streams typed events in real time (plan created, step started, step executed, attestation, feedback request, final response) with keepalive signals to prevent idle timeouts
Pauses execution on human feedback requests, resumes from the exact same state once the user responds
Tracks UID mappings across multi-step operations so subsequent steps can reference resources created in earlier steps
Detects non-recoverable failures (expired sessions, unreachable servers) and aborts gracefully
Persists full conversation state to a database and per-conversation workspace files

Integrated Planning and Execution

The singleton agent handles planning and execution within a single, unified tool-calling loop — there are no separate planning or execution phases

Appendix - Backup and Restore

Mon, 01 Jan 0001 00:00:00 +0000

Backup

The Protegrity Agent Helm chart includes automated Postgres backup and disaster recovery via Velero. Backup runs by default on every installation.

How It Works

A CronJob runs pg_dump every three hours to a dedicated Persistent Volume Claim (PVC). Velero snapshots that PVC to S3 thirty minutes later. This two-stage approach ensures database consistency.

Restore

To recover from a disaster, reinstall the chart with global.restore=true:

helm upgrade --install protegrity-agent \
 --namespace pty-protegrity-agent \
 --set global.restore=true \
 --timeout 15m

To restore a specific point-in-time backup, add --set backup.backupName=<name>.