Your board asks: is that normal? The cloud bill. The team size. The outage count. fsyncDATA Circle is where India's GCC data leaders find out. Numbers stay in. Vendors stay out. Sixty seats, invitation only. fsyncdata.io/request-an-invi… #fsyncDATACircle #GCC

ClickHouse Cloud FinOps: A Board-Ready 12-Metric Scorecard to Stop Overspending chistadata.com/clickhouse-cl… #ClickHouseCloud #FinOps #ClickHouseCloudFinOps #ClickHouseCloudPerformance #ClickHouseCloudPerformance #CloudFinOps #ChistaDATA #Analytics

Supabase FinOps: 10 Essential Metrics to Stop Overspending in 2026 minervadb.com/supabase-finop… #Supabase #SupabaseFinOps #FinOps #CloudOptimization #CloudCost #MinervaDB #Serverless #DBaaS

ClickHouse 26.9 High Availability Architecture chistadata.com/clickhouse-hi… #ClickHouse ChistaDATA Inc. #Analytics #OLAP #ClickHouseConsulting #ClickHouseSupport #ChistaDATA

India has 2,117 GCCs and 2.36M people in them (Nasscom-Zinnov). Many now own the data estate, not the tickets. fsyncDATA is for those who run it. fsyncdata.io #fsyncDATA #GCC #GCCIndia #DataLeaders #CDO #DataArchitecture #DBA #DataEngineering #Database #India

ClickHouse Observability at Scale chistadata.com/clickhouse-ob… #ClickHouse #Observability #Monitoring #ChistaDATA #ClickHouseConsulting #ClickHouseSupport

ClickHouse Cloud Optimization Tips & Tricks chistadata.com/clickhouse-cl… #ClickHouse #ClickHouseCloud #ClickHouseCloudFinOps #ClickHousePerformance #ClickHouseConsulting #ClickHouseSupport #DBA #OLAP #ChistaDATA

Postgres Lakehouse Integration on Snowflake minervadb.com/postgres-lakeh… #Postgres #Snowflake #PostgreSQL #Analytics #DataLake #MinervaDB

ClickHouse 26.9 High Availability improvements - chistadata.com/clickhouse-hi… #ClickHouse #ClickHouseHighAvailability #ClickHouseConsulting #ClickHouseSupport #ChistaDATA #OLAP #Analytics

For years I told people a composite index won't help if the query skips the leading column. With skip scan in Postgres 18 that's not always true anymore, as long as the leading column has few distinct values. Check your plans before adding one more index. #PostgreSQL #PostgreSQLPerformance #PostgreSQLTips #PostgreSQLConsulting #PostgreSQLTips

This is great, @BenjDicken . Your interactive pieces on B-trees and disk I/O taught a lot of people more than most textbooks do, and this one does the same for hardware almost none of us will ever touch. From the database side, the part I keep wanting to understand better is the memory path between the Grace CPU and the GPU. That's where a lot of the "move compute to the data" ideas start to look different. Bookmarked.

Tuning YugabyteDB for optimal performance minervadb.com/yugabytedb-per… #Yugabyte #YugabyteDBPerformance #MinervaDB #ConsultativeSupport #SQL #DataPerformance #YugabyteDBConsulting #YugabyteDBSupport

An ultimate Database Infrastructure Partner for your business - minervadb.com/database-infra… #MinervaDB #SQL #NoSQL #NewSQL #DatabaseSystems #OpenSource #DBaaS #Cloud #MinervaDBPartnership #DataAnalytics #OLTP #OLAP #Consulting #DBASupport #RemoteDBA #CloudNative #DataScience #AI #Data #DataArchitect #DataEngineering

Understanding Query Execution inside ClickHouse chistadata.com/clickhouse-qu… #ClickHouse #ChistaDATA #ClickHouseConsulting #ChistaDATA #ClickHouseSupport #ClickHousePerformance #Analytics #DBA #ClickHouseTips

Building Observability and Monitoring for MongoDB minervadb.com/mongodb-observ… #MongoDB #MongoDBPerformance #MongoDBObservability #MongoDBMonitoring #MongoDBSupport #MongoDBDBA #MongoDBConsulting #Monitoring

Scaling Valkey? My order: Enable io-threads before adding shards, if you have spare cores. Then hunt hot keys and slow Lua. More threads won't fix those. Only then reshard, with atomic slot migration on 9.x. Test on your own workload. #Valkey #Redis #Scalability

GCCs on Databricks get asked one thing by the parent company: what does each team cost? The data is already in the account, it just isn't attributed yet. Source of truth is system.billing.usage, which gives DBUs by sku_name with usage_metadata (cluster_id, job_id, warehouse_id) and identity_metadata.run_as. Join system.billing.list_prices on sku_name and the price window to get cost, but remember that's list price, not your contract rate. Attribution needs tags you can trust, so use cluster policies with custom_tags.cost_center as fixed or allowlist and nobody launches an untagged cluster. SQL warehouses carry their own tags, and serverless has no cluster to tag, so attribute it through budget policies. Classic compute also has two bills: the DBUs sit in system tables, the VMs sit in your cloud bill. Custom tags propagate to the cloud resources, so the same cost_center key joins both. Skip that and your chargeback undercounts. Then the obvious ones: scheduled work on jobs compute, not all-purpose, and autotermination_minutes capped in the policy. Billing tables lag by hours, so this is chargeback, not real-time alerting. Test on your own workload. #Databricks #FinOps #GCC

Went through Snowflake's 2026 performance notes. Most of it is quiet planner and pruning work: TIMESTAMP_TZ pruning, skew-aware hash joins, filters pushed below CTEs, worker reuse for short queries. Optima Planning is Gen2 only. Test on your own workload. #Snowflake #SQL #MinervaDB

Understanding Snowflake and Databricks FinOps for Performance, Reliability and Cost - minervadb.com/snowflake-and-… #snowflake #MinervaDB #Databricks @WebScaleDBA #fsyncDATACircle #GCC #Data #Analytics #DataScience #OLAP

An invitation to GCC leaders: what the fsyncDATA GCC programme gives you fsyncdata.io/gcc-leaders-inv… #fsyncdata #fsyncdatagcc #GCC #Leadership #AI #Data #Analytics