Host Observability

The Coralogix Host integration uses OpenTelemetry to collect logs, metrics, and traces from Linux hosts, including virtual machines, bare-metal servers, and cloud instances.

The integration deploys an OpenTelemetry Collector directly on the operating system as a long-running service. The agent collects telemetry from local applications, OS services, and the host, and exports it to Coralogix.

Host observability is machine-level. Configuration changes apply only to the host where the agent is installed. Installation, upgrades, permissions, and restarts are managed using native operating system tools.

Integration overview

The Host integration enables deep observability into applications and infrastructure running directly on a Linux system:

• Application monitoring - Collect traces, generate metrics from spans, apply sampling strategies, and enable profiling for applications running on the host—without requiring containerization.
• Host and service logs - Ingest logs directly from the systemd journal, OS services, and application log files. Apply filtering, multiline parsing, and source-level volume control before logs are sent downstream.
• Infrastructure and system metrics - Capture CPU, memory, disk, network, and process-level metrics to understand the health and performance of the host or VM itself.
• Agent visibility and management - Monitor the OpenTelemetry Collector as a local process, with optional centralized configuration and monitoring through Fleet Management (OpAMP).

Coralogix features enabled

Unlock deep observability and cost-efficient monitoring with a single integration—no manual pipeline construction required.

• APM - End-to-end tracing using OpenTelemetry for applications running directly on the host. Supports head and tail sampling strategies and automatically generates span metrics to ensure full APM visibility.
• Continuous Profiling - Analyze application behavior over time by collecting CPU usage, execution paths, and performance characteristics directly from host-based workloads.
• Logs and metrics - Collect enriched system logs, service logs, application logs, and host metrics, all tagged with host-level metadata for filtering, correlation, and analysis across the Coralogix platform.
• **Infrastructure Explorer** - Explore a live inventory of hosts and virtual machines with real-time context, including resource usage, ownership, and relationships.
• Fleet Management - Centrally monitor and manage OpenTelemetry agents running on individual hosts using OpAMP, enabling scalable configuration and version control across your fleet.

Logs configuration

Host environments support broad and flexible logging model.

System journal logs (journald)

The integration can ingest logs directly from the systemd journal using the journald receiver. This provides access to structured system logs, including metadata such as service name, process ID, and priority.

The agent requires permission to read the system journal. Ensure the agent user belongs to the appropriate system groups.

Journal filtering

Filter logs at the source using journal match expressions to reduce noise and control volume:

• Filter by systemd unit
• Include or exclude specific services
• Apply priority-level filtering

Filtering at the journal level prevents unwanted logs from being ingested altogether.

File-based log collection

For applications that write logs to disk, the integration supports file-based log receivers.

• Specify explicit file paths or glob patterns
• Automatically handle log rotation
• Configure multiple file log receivers per host

Example patterns:

• /var/log/*.log
• /opt/myapp/logs/**/*.log

Multiline log parsing

Host logs often contain stack traces or multiline entries. Enable multiline parsing to correctly group related log lines:

• Define a regex pattern that identifies the start of a new log entry
• Commonly based on timestamps or severity markers

Prerequisites

This integration is designed and tested for clean installs. Applying it to an existing OpenTelemetry Collector installation may require additional validation and manual adjustments.

Before installing the integration, ensure the following:

• Linux host with systemd - A supported distribution (Debian/Ubuntu for .deb, RHEL/Amazon Linux for .rpm) and an available package manager (apt, dnf, or yum).
• Network access - Outbound HTTPS (TCP 443) to Coralogix endpoints, with local ports 4317, 4318, and 8888 available (or custom ports if configured).

• File system and security permissions - Read access to /var/log and the systemd journal, and write access to /etc/otelcol or /etc/otelcol-contrib.

  If SELinux is enforcing, allow the agent to bind to the configured ports

Installation

Follow the steps below to deploy the Coralogix Host Observability integration on a Linux machine.

Enable the Use AWS PrivateLink option to configure your integration to use a PrivateLink endpoint.

Select observability features

1. Choose the observability features you want to enable. Toggling a feature updates the generated configuration file accordingly.
2. Click Save to generate the host-specific configuration.

Select installation method

The Host integration supports multiple installation methods depending on your environment and operational preferences:

• Installer script (recommended):
  Automatically detects system architecture and installs the appropriate package.

• Package-based installation:
  Install the agent using a native package format:

• DEB (Debian / Ubuntu)

• RPM (RHEL / CentOS / Amazon Linux)

• Standalone binary:
  Manually download and run the collector binary.

Ensure you select the correct package format and system architecture (AMD64 or ARM64). Installing an incompatible binary will prevent the agent from starting.

Limitations

• Coralogix enforces a hard limit of 10MB per payload to OpenTelemetry ingestion endpoints, with a recommended maximum of 2MB.
• Metric names must be a maximum of 255 characters.
• Attribute keys for metric data must be a maximum of 255 characters.

Next steps

• Learn how to scale host configuration using Fleet Management.

Additional resources

Support

Need help?

Our world-class customer success team is available 24/7 to walk you through your setup and answer any questions that may come up.

Reach out via in-app chat or email us at [email protected].