Aviz details ONES telemetry scale test results

Aviz reports that its ONES telemetry collectors were evaluated under scale tests and sustained low, consistent latency while supporting large device counts, using a gNMI simulator and Datadog integration to validate performance and resource use.

Research overview

The evaluation focused on how ONES behaves when telemetry collectors are exercised across increasing scale and operational stressors. The testing program combined architectural adjustments, resource management checks, and iterative load validation to assess collector behavior.

Key findings

Under scale conditions that simulated more than one thousand devices, ONES maintained consistent low latency for ingest and write operations. Data integrity checks showed alignment between source of truth and delivered telemetry, and fault tolerance mechanisms preserved collection during simulated disruptions.

Technical breakdown

Engineers used an internal gNMI simulator to generate large volumes of telemetry traffic and concurrent connections, enabling assessment of parsing, buffering, and write paths. Resource telemetry was monitored with Datadog to track Central Processing Unit (CPU), memory, and storage use and to guide configuration adjustments.

Latency measurement

Latency was measured continuously while the collectors handled a scale scenario of approximately 1,024 devices and showed consistent low processing times from receipt to database push.

Operational impact

Resource optimization efforts reduced bottleneck risk by tuning collector configurations and lifecycle behavior, supporting sustained operation over extended runs. Traffic-control tests in a Linux environment validated collector responses to network congestion and packet disruption scenarios.

Testing tools and methods

Data integrity tests were applied across ingestion and delivery stages to compare delivered metrics against the source of truth. The scale simulator architecture combined traffic generation, observability telemetry, and failure injection to reveal capacity limits and recovery paths.

The summary above presents verifiable results from the vendor blog and is intended for technical decision-makers evaluating telemetry solutions. This “Blog Signals brief” is a fact-based summary of the vendor blog.