Skip to content

Installation

Applies to1.1.0DifficultyBeginner

This page walks through connecting Oculis to your infrastructure for the first time. For the complete installation reference — every platform, prerequisite, health check, and troubleshooting step — see Oculis Collector.

Oculis Collector is the component you install on infrastructure running AI workloads. It discovers supported AI workloads and reports operational telemetry to Oculis.

It runs out-of-band: it observes the host it is installed on and sends data outbound to Oculis. It does not sit in the model request path, so a Collector outage does not interrupt your application or model traffic.

You will need:

  • Access to Oculis
  • Administrative access to the host you are installing on
  • Outbound HTTPS connectivity from that host to Oculis
  • A supported Linux, macOS, Windows, or Docker environment
  1. Sign in to Oculis.

  2. Choose Add Collector.

  3. Select the platform you are installing on.

  4. Copy the generated installation command.

  5. Run it on the intended host.

  6. Wait for the Collector to appear as connected in Oculis.

Platform-specific commands and prerequisites are listed on the Oculis Collector page.

Once the Collector is connected:

  • It begins discovering supported AI workloads on the host.
  • Supported host instrumentation is configured where applicable.
  • Workloads appear in Oculis as they run, so a host with idle applications may take a little time to populate.
  • Application telemetry becomes available once a workload has the required Oculis instrumentation.

The Collector appears as connected in Oculis when it has successfully reported in. That is the simplest confirmation that installation worked.

For platform-specific health checks and troubleshooting, see Oculis Collector.

The Collector can be deployed on:

  • Linux
  • macOS
  • Windows
  • Docker

Kubernetes and Helm deployment are not supported in Oculis Collector 1.1.0.