10.2 Monitoring

Using OpenTelemetry in Thinfinity® Workspace

OpenTelemetry is the standard adopted by Thinfinity Workspace for the collection, processing, and exporting of observability signals. These signals enable the analysis of the components that are part of the Thinfinity Workspace deploy network, by focusing on two main actions:

• Monitoring: Identifying problems using predefined metrics, such as CPU or memory usage.

• Observability: Analyzing the system's status based on metrics and logs collected from multiple sources, providing a deeper insight into how applications are functioning.

Signal Collection and Export

With OpenTelemetry, two types of signals are collected:

• Metrics These are temporal data that reflect the system's status and performance over time. Thinfinity Workspace uses distributed agents, such as Node Collector and Gateway Collector, to collect and process these metrics, which are then exported to an observability backend with a TSDB (Time Series Database), with VictoriaMetrics being the default solution.

• Logs These are timestamped records that document system events. Thinfinity Workspace uses the Enterprise Logger Service component to consolidate logs from all components (Brokers, Gateways, Agents). Logs are exported in OpenTelemetry format, and although a default log backend is not included, export to compatible solutions is supported. VictoriaLogs is an example of a backend that can be integrated for visualization and analysis.

Transfer Protocol

Thinfinity Workspace utilizes the OpenTelemetry Protocol (OTLP) standard for collecting, processing, and exporting observability signals, which primarily consist of metrics and logs, offering insight into system and application behavior.

For transmitting this telemetry data, Thinfinity Workspace supports two OTLP transport modes: OTLP/HTTP and OTLP/gRPC. Specifically.

• OTLP/HTTP is used for metric transfer between Node Collectors (from Brokers) and the Gateway Collector, and is a configurable option for the final export of both metrics and logs from the Gateway Collector to the Observability Backend. OTLP/HTTP is also supported for integrating with external backends like VictoriaLogs (for logs), Prometheus (for metrics), and Dynatrace (for both).

• OTLP/gRPC is specifically employed by the Enterprise Logger Service to send consolidated logs to the Gateway Collector, and it is also a configurable option for the final export of both metrics and logs from the Gateway Collector to the Observability Backend.

Both OTLP configurations are managed via the Telemetry section in the Thinfinity® Workspace Configuration Manager.

Reference

For a better understanding of the Telemetry tab Interface and how to use it, see Telemetry Tab.

Visualizing Metrics with Observability Backends

Observability Backends is a third-party platform used to manage the database where all metrics will be stored and imported.

Thinfinity Workspace offers an integrated Monitoring interface, accessible from the user profile icon ("Monitoring" option), where real-time and historical metrics can be consulted. The metrics visualized in this interface come from VictoriaMetrics, the backend that also allows for advanced analysis if its database is accessed directly.

Monitoring allows data collection from the following components:

• Gateways: Shows the CPU and memory usage for the Gateway process and the host.

• Brokers: Shows CPU and memory usage for the broker process and the host.

• Services: Presents similar metrics for other platform services.

• Agents: Shows CPU and memory usage for the Agent process.

• Connections: Reports on the duration, type, and quantity of active connections.

• Sessions: Details active sessions, indicating user, date, and start time.

Reference

For a better understanding of the Monitoring Interface and how to use it, see Monitoring.

Visualizing Logs

For logs, visualization depends on integration with external OpenTelemetry compatible backends, such as VictoriaLogs.

Audit Log

The Enterprise Logger Service consolidates logs from all components into a single data stream. This function allows for detailed tracking of system activity: access, sessions, configuration changes, etc. Each recorded event includes information such as: component, origin, user, IP, result, and contextual message.

Integration with VictoriaLogs

Although Thinfinity Workspace does not include a backend for logs, it allows exporting them to VictoriaLogs. Once configured (from the Configuration Manager, by setting the URL and protocol), logs can be visualized in their web interface:

Access to the interface: [protocol]://[GatewayURL][:port]/select/vmui

For more advanced configuration, it is recommended to consult the official VictoriaLogs documentation.

Monitoring works with VictoriaLogs and also supports exporting metrics to backends such as Prometheus, Splunk, Elastic Stack, Datadog, Azure, and Dynatrace.