Once you have built and tested your automation, you can deploy your bot to a live environment. Narakim manages the full deployment lifecycle — from containerization to orchestration to health monitoring.

How Deployment Works

You build and validate your automation on the Logic canvas.
You navigate to the Deployments tab and create a deployment configuration.
You click Deploy — Narakim packages your automation graph, resolves secrets, and spins up an isolated Docker container running the Automation Engine.
The bot instance starts processing events in real time.
You monitor the instance status, metrics, and log stream from the Studio.

Deployment Configuration

Setting	Options	Description
Environment	`DEVELOPMENT`, `PRODUCTION`, `DEVELOPMENT`	Logical grouping for the deployment
Provider	`LOCAL_DOCKER`, `AWS_ECS`, `AWS_LAMBDA`, `AZURE_CONTAINER`, `GCP_CLOUD_RUN`	Where the container runs
Resource Tier	`SMALL`, `MEDIUM`, `LARGE`, `XLARGE`	Compute and memory allocation
Health Check	Path + interval	Endpoint to verify the container is healthy
Auto-Restart	Enabled / Disabled	Automatically restart the container on crash
Auto-Scale	Enabled / Disabled	Scale horizontally based on load (provider-dependent)
Env Vars	Key-value pairs	Custom environment variables injected at runtime

Bot Instance Lifecycle

STOPPED → STARTING → RUNNING
                         │
                    ┌────┴────┐
                    ▼         ▼
                  ERROR    DEGRADED
                    │
                    └── STOPPED (on auto-restart)

State	Meaning
STOPPED	The bot instance is not running.
STARTING	Container is being provisioned and initialized.
RUNNING	Bot is live and processing events.
ERROR	A fatal error has occurred. Check the log stream.
DEGRADED	Partial functionality — some extensions may be failing.

Monitoring a Running Instance

From the Deployments tab, you can see:

CPU and Memory usage — Live metrics for the running container.
Uptime — How long the current instance has been running.
Last Ping — Timestamp of the last health check response.
Activity Timeline — Deployment events: started, progress updates, completed, failed.
Log Stream — Real-time INFO, WARN, ERROR, and DEBUG entries from the Automation Engine.

Stopping a Bot

Click Stop on the running instance in the Deployments tab. The container is gracefully shut down. You can restart it at any time by clicking Deploy again.

Development vs Production

Use Development for testing with non-production data. Secrets can be scoped to an environment — configure separate development and production secrets in the Configurations tab to avoid using live credentials during development.

Was this helpful?

historyLast updated: May 10, 2026