--- title: How LinkedIn extended ClickHouse from distributed... description: LinkedIn's observability team describes how it extended its ClickHouse footprint beyond distributed tracing into metric discovery and analytics. After proving... canonical: https://daily.dev/posts/how-linkedin-extended-clickhouse-from-distributed-tracing-to-metric-discovery-and-analytics-6fy7jqf46 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: How LinkedIn extended ClickHouse from distributed tracing to metric discovery and analytics | daily.dev og:description: LinkedIn's observability team describes how it extended its ClickHouse footprint beyond distributed tracing into metric discovery and analytics. After proving... og:url: https://daily.dev/posts/how-linkedin-extended-clickhouse-from-distributed-tracing-to-metric-discovery-and-analytics-6fy7jqf46 og:image: https://api.daily.dev/og/posts/6FY7JqF46.png og:image:alt: How LinkedIn extended ClickHouse from distributed tracing to metric discovery and analytics og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # How LinkedIn extended ClickHouse from distributed tracing to metric discovery and analytics **[ClickHouse](https://daily.dev/sources/clickhouse)** · 10 min read · 4 upvotes · 0 comments ## Summary LinkedIn's observability team describes how it extended its ClickHouse footprint beyond distributed tracing into metric discovery and analytics. After proving ClickHouse at scale with a tracing system handling roughly 800 billion spans a day across three data centers, the team consolidated a legacy metric metadata stack built on RRDtool-style naming, a custom in-memory search index, a custom in-memory KV store, and Elasticsearch into a single ClickHouse index using a ReplacingMergeTree table and projections for heavy aggregations. The new system serves over 150,000 queries per minute at 68ms average latency across 13 billion-plus metrics, cut memory usage to roughly one-fifth and compute to about two-thirds compared to the old stack, and leaves headroom to double metric volume on a single shard. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How many queries per minute does LinkedIn's ClickHouse-based metric metadata index serve and at what latency? The consolidated ClickHouse index serves more than 150,000 queries per minute across LinkedIn's fleet with an average latency of 68 milliseconds, covering over 13 billion metrics with roughly 30% daily turnover, all handled on a single shard. The average includes both fast lookups and slow analytical queries that older systems could not serve at all. _Engineers evaluating ClickHouse for metadata workloads can follow real production benchmarks like this on daily.dev._ ### What resource savings did LinkedIn achieve by migrating its legacy metric metadata stack to ClickHouse? Migrating off the legacy three-system metric metadata stack to ClickHouse cut memory usage to roughly one-fifth and reduced compute to about two-thirds of previous requirements. The new single-shard index also has enough headroom to double metric volume without resharding, a significant improvement over the old architecture, which had already reached its scaling limits. _Teams weighing infrastructure consolidation costs can track real-world migration outcomes like this on daily.dev._ ### Why did LinkedIn choose ClickHouse over Elasticsearch or a custom in-memory solution for metric metadata? ClickHouse was chosen because it supported all existing queries faster than the custom in-memory databases it replaced, while Elasticsearch had already failed an earlier migration attempt and struggled with large spanning aggregations and in-memory serialization of big result sets. Building another custom solution was rejected as inflexible and reliant on limited in-house expertise. _Developers comparing ClickHouse, Elasticsearch, and custom stores for analytics can weigh trade-offs like these on daily.dev._ ## Similar posts on daily.dev - [How the D. E. Shaw group powers high-cardinality observability at scale with ClickHouse](https://daily.dev/posts/how-the-d-e-shaw-group-powers-high-cardinality-observability-at-scale-with-clickhouse-51pw0hsby) · ClickHouse · 0 upvotes · 0 comments - [How Jaeger hit 8.6× compression on 10 million spans with ClickHouse](https://daily.dev/posts/how-jaeger-hit-8-6-compression-on-10-million-spans-with-clickhouse-n5yas4v4u) · The New Stack · 1 upvotes · 0 comments - [Building Jaeger’s ClickHouse backend: 8.6× compression on 10 million spans](https://daily.dev/posts/building-jaeger-s-clickhouse-backend-8-6-compression-on-10-million-spans-3dmlhztxi) · CNCF · 3 upvotes · 0 comments - [Have you heard? Clickhouse is winning the observability wars\!](https://daily.dev/posts/have-you-heard-clickhouse-is-winning-the-observability-wars--5bpefin3l) · Charity · 18 upvotes · 0 comments - [How we made ClickStack 5x faster for ClickHouse observability](https://daily.dev/posts/how-we-made-clickstack-5x-faster-for-clickhouse-observability-sr5nkmvis) · ClickHouse · 3 upvotes · 0 comments --- Tags: [#devops](https://daily.dev/tags/devops), [#observability](https://daily.dev/tags/observability), [#elk](https://daily.dev/tags/elk), [#linkedin](https://daily.dev/tags/linkedin), [#clickhouse](https://daily.dev/tags/clickhouse) [View this post on daily.dev](https://daily.dev/posts/how-linkedin-extended-clickhouse-from-distributed-tracing-to-metric-discovery-and-analytics-6fy7jqf46) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"How LinkedIn extended ClickHouse from distributed tracing to metric discovery and analytics","url":"https://daily.dev/posts/how-linkedin-extended-clickhouse-from-distributed-tracing-to-metric-discovery-and-analytics-6fy7jqf46","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/how-linkedin-extended-clickhouse-from-distributed-tracing-to-metric-discovery-and-analytics-6fy7jqf46"},"datePublished":"2026-10-01T11:49:42.114Z","dateModified":"2026-10-01T12:10:12.595Z","description":"LinkedIn's observability team describes how it extended its ClickHouse footprint beyond distributed tracing into metric discovery and analytics. After proving...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5935723f795b9376c452bbc938fbce1d?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5935723f795b9376c452bbc938fbce1d?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"ClickHouse","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"ClickHouse","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/clickhouse","url":"https://daily.dev/sources/clickhouse"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/how-linkedin-extended-clickhouse-from-distributed-tracing-to-metric-discovery-and-analytics-6fy7jqf46","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":4},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"devops,observability,elk,linkedin,clickhouse","timeRequired":"PT10M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"ClickHouse","item":"https://daily.dev/sources/clickhouse"},{"@type":"ListItem","position":3,"name":"How LinkedIn extended ClickHouse from distributed tracing to metric discovery and analytics"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/how-linkedin-extended-clickhouse-from-distributed-tracing-to-metric-discovery-and-analytics-6fy7jqf46#faq","mainEntity":[{"@type":"Question","name":"How many queries per minute does LinkedIn's ClickHouse-based metric metadata index serve and at what latency?","acceptedAnswer":{"@type":"Answer","text":"The consolidated ClickHouse index serves more than 150,000 queries per minute across LinkedIn's fleet with an average latency of 68 milliseconds, covering over 13 billion metrics with roughly 30% daily turnover, all handled on a single shard. The average includes both fast lookups and slow analytical queries that older systems could not serve at all. Engineers evaluating ClickHouse for metadata workloads can follow real production benchmarks like this on daily.dev."}},{"@type":"Question","name":"What resource savings did LinkedIn achieve by migrating its legacy metric metadata stack to ClickHouse?","acceptedAnswer":{"@type":"Answer","text":"Migrating off the legacy three-system metric metadata stack to ClickHouse cut memory usage to roughly one-fifth and reduced compute to about two-thirds of previous requirements. The new single-shard index also has enough headroom to double metric volume without resharding, a significant improvement over the old architecture, which had already reached its scaling limits. Teams weighing infrastructure consolidation costs can track real-world migration outcomes like this on daily.dev."}},{"@type":"Question","name":"Why did LinkedIn choose ClickHouse over Elasticsearch or a custom in-memory solution for metric metadata?","acceptedAnswer":{"@type":"Answer","text":"ClickHouse was chosen because it supported all existing queries faster than the custom in-memory databases it replaced, while Elasticsearch had already failed an earlier migration attempt and struggled with large spanning aggregations and in-memory serialization of big result sets. Building another custom solution was rejected as inflexible and reliant on limited in-house expertise. Developers comparing ClickHouse, Elasticsearch, and custom stores for analytics can weigh trade-offs like these on daily.dev."}}]} ```