VictoriaLogs provided Airwallex with the best balance of performance, simplicity, and cost.

  • FinTech
  • Singapore

VictoriaLogs provided a sustainable cost model for large-scale log ingestion. Compared with more customizable or warehouse-style architectures, it was much faster to operationalize and required less platform engineering overhead.

Main Benefits of Using VictoriaLogs

  • Savings icon representing cost effectiveness

    Cost-effectiveness

  • Rocket icon representing ingestion efficiency

    High ingestion efficiency

  • Speed icon representing query performance and reliability

    Strong query performance and reliability

  • Settings icon representing operational simplicity

    Operational simplicity

  • Integrations icon representing stack compatibility

    Compatibility with the observability stack

  • Scalability icon representing growth at scale

    Scalability

Challenge

Before adopting VictoriaLogs, Airwallex was facing a combination of cost, scale, and reliability challenges in its logging stack.

As Airwallex’s platform and engineering organization grew, log volume increased significantly across Kubernetes workloads, infrastructure components, and cloud-native services. Airwallex’s existing setup relied heavily on premium log platforms and cloud-native logging backends, which created rising ingestion costs and increasing pressure to control spend without sacrificing operational visibility.

In short, Airwallex needed a solution that could efficiently absorb large log volumes, scale with their business, materially reduce costs, and still provide a good developer/operator experience for troubleshooting and alerting.

At the same time, Airwallex needed to address several technical issues:

  • Risk of throttling during peak ingestion periods on the primary logging platform
  • A growing need to separate high-value operational/security logs from lower-value or high-volume logs
  • Limited flexibility in managing logs from multiple environments and regions in a cost-efficient way
  • The need for a simpler, more scalable logging backend for broad engineering self-service
  • The need to preserve telemetry reliably while reducing dependence on expensive systems for all log categories

Solution

VictoriaLogs allowed Airwallex to position their logging platform more deliberately: premium systems for the highest-value, most time-sensitive logs, and VictoriaLogs for broader, large-scale retention and exploration use cases.

  • High-volume log ingestion at lower cost: it gave Airwallex a practical way to offload large volumes of low-value, high-cardinality logs from premium platforms.
  • Simple architecture and ease of adoption: compared with more operationally intensive alternatives, VictoriaLogs was much simpler to deploy and operate.
  • Good query experience: the built-in query capabilities and Grafana integration made it easy for engineers to explore logs without introducing a completely new workflow.
  • Compatibility with Airwallex’s observability stack: it fits well with their OpenTelemetry-based ingestion approach and broader observability architecture.
  • Operational efficiency: the platform was simple enough that the Airwallex team could scale usage without creating a large operational burden.

Why VictoriaLogs Was Chosen Over Other Solutions

  • Training icon representing fast production rollout

    VictoriaLogs provided a fast path from evaluation to production use.

  • Rocket icon representing strong ingestion and query performance

    Strong ingestion and query performance for Airwallex’s target workloads

  • Agile icon representing lower operational complexity

    Lower operational complexity than more custom alternatives.

  • Integrations icon representing native workflow fit

    Native fit with Grafana and OpenTelemetry-based workflows.

  • VictoriaLogs offered better economics for high-volume logs while preserving a practical day-to-day experience for engineering teams.

Technical Stats

  • Airwallex Logging Setup

    Running the open source VictoriaLogs Cluster version

  • Platform

    With a dedicated nodepool on Google Kubernetes Engine

  • Log retention

    1 month

  • Ingestion Rate

    sum(rate(vl_rows_ingested_total[24h]))

    ~1 Million/day

  • Disk Usage (Data)

    sum(vl_data_size_bytes{type="storage"})

    ~20TB for 1 Month retention

  • Disk Usage (Index)

    sum(vl_data_size_bytes{type="indexdb"})

    ~3.5TB for 1 Month retention

  • Log Search Rate

    sum(rate(vl_http_requests_total{path=~".*/select/.*"}[24h]))

    ~20

  • Memory Usage

    process_resident_memory_bytes

    145GB

  • CPU Usage

    rate(process_cpu_seconds_total[24h])

    150

What Airwallex Had to Say

  • “VictoriaMetrics products are high-performance, operationally efficient building blocks for modern observability. They are designed to handle scale well, integrate cleanly with common open-source tooling, and give platform teams a practical way to control cost without giving up visibility. In our experience, VictoriaLogs has been especially valuable as a simple, scalable log platform for high-volume use cases.”
    Luxin Chen, Senior Software Engineer, Airwallex