Managing Deployments in Agiloop

Agiloop's deployment management tools let you provision cloud infrastructure, monitor running services, and inspect logs — all from within the platform. This article walks through the full deployment lifecycle, from creating a deploy project to tearing it down.

Creating a Deploy Project

Agiloop can deploy your generated apps and services to a hosted cloud infrastructure service. Agiloop currently supports Railway. When you provide your own Railway account API key, IMPLEMENT will generate a deployment to Railway for each service. This will happen automatically.

The Deployment Manifest

The deployment manifest describes everything Agiloop needs to run your application. It is made up of three sections and is generated automatically by the build system.

• Services: The application containers that will be built and run (for example, a web server or an API).
• Databases: Managed database instances attached to your environment (such as PostgreSQL or Redis).
• Resources: Additional cloud resources your application depends on (such as storage buckets or queues).

Setting Environment Variables

Environment variables are configured at each service.

Service-Level Variables

Variables set at the service level apply only to that specific service. Use these to provide configuration that is unique to one service (for example, a worker queue connection string that only the background service needs).

To add or update a variable, navigate to the relevant service or the project-level settings within your deploy project, then add the key-value pair.

Note: Changes to environment variables typically require a redeployment to take effect. Agiloop will prompt you if a redeploy is needed.

Viewing Deployment Status

Once a deploy is running, the deployment status view gives you a live overview of your environment. From here you can see which services are running, pending, or in an error state, and quickly identify any issues that need attention.

Inspecting Logs

Agiloop provides four distinct log types so you can trace exactly what happened at every stage:

• Build logs: Output from the container image build process — compiler output, dependency installs, and build errors.
• Deploy logs: Events from the deployment orchestrator as it rolls out new containers and routes traffic.
• HTTP logs: Incoming request logs for your services, showing method, path, status code, and latency.
• Orchestration logs: Lower-level platform events related to scheduling, health checks, and infrastructure changes.

Real-Time Log Streaming

All four log types support real-time streaming. When you open a log view, new entries appear automatically as they are generated, without needing to refresh the page. This is especially useful when monitoring an active deployment or debugging a service that is misbehaving under live traffic.

To stream logs:

1. In your deploy project, select the service you want to inspect.
2. Choose the log type from the available tabs (Build, Deploy, HTTP, or Orchestration).
3. Logs stream in automatically. Scroll up to review historical entries or stay at the bottom to follow the live tail.

Related Articles

• Deploying Your App

• Git Repository Configuration

• Feature Review Process

• Code Generation Skills

• Project Settings: A Complete Guide