When evaluating Apache Doris(https://doris.apache.org/), Elasticsearch, or ClickHouse for observability, you're really deciding how to handle massive volumes of fast-moving, constantly evolving data. Four questions teams should ask: 1️⃣ How much will it cost to store all this data? → Elasticsearch: Storage-heavy indexes, most expensive. → ClickHouse: Good compression, may require more tuning. → Apache Doris: Very high compression, 50–80% cheaper than Elasticsearch. Also offers storage-compute separation, hot data on cloud disks and cold data in object storage. 2️⃣ Can it ingest data in real time? → Elasticsearch: slows under high throughput ingest → ClickHouse: strong ingest → Apache Doris: 10 GB/s real-time ingest, handling PB-scale observability data daily 3️⃣ Can it search text fast and run complex analytics? → Elasticsearch: built its name in full-text search, slower analytics → ClickHouse: good in analytics, text search still experimental → Apache Doris: great in analytics and full-text search. Offer inverted indexes + columnar engine → 3–10x faster full-text search than ClickHouse and 6–21x better aggregation performance than Elasticsearch. 4️⃣ Will the schema break as logs evolve? → Elasticsearch: uses dynamic mapping, but type conflicts are painful → ClickHouse: schema changes require planning → Apache Doris: provides flexible schema with VARIANT data type, supports changing field type as data changes and large-scale JSON analytics. 🔗 See demo on OpenTelemetry + Apache Doris + Grafana: https://lnkd.in/geS-WNty

Ups, seems Matomo is not setup properly on the page

This looks excellent thank you for sharing. 🙂

I opened a thread in the doris slack, let's see 🙂

Cool thank you Sven

sorry to reply late. at that time, we have urgently fixed the issue with the official website😆

Yeah, thx - I received the answer in the thread on Doris slack at Tuesday :)