This post describes how the author built a logging platform with ClickHouse, which saved them millions of dollars compared to using Datadog. They explain the challenges they faced with Datadog and the decisions they made, such as choosing OpenTelemetry and enhancing Grafana. They also provide cost analysis, performance results, and future plans.
Table of contents
Table of ContentsIntroductionA small team with an enormous challengeChoosing ClickHouseThe scale of the problemClickHouse Cloud infrastructureBuilding on top of ClickHouse CloudUnderstanding your usersSome early design decisionsChoosing OpenTelemetry (OTel)Architectural overviewIngest processingSchemaEnhancing GrafanaCross-region queryingPerformance and costLooking forwardConclusion3 Impressions