Why We Chose the M3DB Data Store for Logz.io Prometheus-as-a-Service

Logz.io has focused on enhancing its observability service by transitioning from using solely Elasticsearch for metrics storage to exploring alternative databases compatible with Prometheus to address cost and performance issues. Initially reliant on Elasticsearch, Logz.io faced challenges with high storage costs and suboptimal performance due to Elasticsearch's unsuitability for time-series data. To better accommodate users already utilizing Prometheus and to streamline cloud migration, Logz.io evaluated four options: VictoriaMetrics, M3DB, Thanos, and Cortex, considering factors like scalability, cost, and integration with multi-region cloud setups. Ultimately, M3DB, an open-source project from Uber, was selected for its full compatibility with Prometheus, despite some limitations in its community diversity and robustness. Logz.io has committed to contributing to and scaling M3DB to support its Prometheus-as-a-Service offering, leveraging Kubernetes for deployment and management.