Overview
Ingestion is how data reaches your Context Lake. You can ingest data using native integrations, MCP connectors, APIs, webhooks, and custom integrations.
This page helps you choose the right ingestion method. For a conceptual explanation of the Context Lake, see Context Lake overview.
How to choose an ingestion method
Use these guidelines as a starting point:
- Choose a native integration when the source system is a clear source of truth (for example, GitHub repositories, PagerDuty services, Kubernetes workloads), and you want continuous sync.
- Choose an MCP connector when you want Port AI and agents to query external tools at runtime (for example, Notion, Linear, Slack), without ingesting all data into the catalog.
- Choose ingestion via API when you already have a system that can push structured data to Port, or when you need precise control over updates.
- Choose ingestion via webhooks when you want event-driven updates from your systems, or when your source can emit events but does not have a suitable integration.
- Choose a custom integration when you need a reusable ingestion component for a tool that Port does not support with a native integration yet.