#oltp
Another database business question is whether an ISV should start with an OLAP DBMS then add OLTP, or start with an OLTP DBMS then add OLAP.
‣ ClickHouse, Snowflake, Databricks are doing OLAP→OLTP.
‣ Supabase, MongoDB, PlanetScale(?) are doing OLTP→OLAP.
November 11, 2026 at 5:41 PM
🌍 Oracle ACEs, join us today as we explore always-on multicloud OLTP, AI & analytics with Globally Distributed AI Database.

🎙️ Deeksha Sehgal
🕖 Nov 18, 7–8 AM PT
#OracleACE
November 18, 2026 at 10:15 AM
[速報]PostgreSQL互換のサーバレスでOLTPとOLAP両対応を実現する「Microsoft Fabric HorizonDB」パブリックプレビュー
https://www.publickey1.jp/blog/26/postgresqloltpolapmicrosoft_fabric_horizondb.html PulickeyPost
[速報]PostgreSQL互換のサーバレスでOLTPとOLAP両対応を実現する「Microsoft Fabric HorizonDB」パブリックプレビュー
マイクロソフトは日本時間11月18日未明に開幕した年次イベント「Microsoft Ignite 2026」で、PostgreSQL互換のサーバレスデータベース「Microsoft Fabric HorizonDB」のパブリックプレビュー版...
www.publickey1.jp
November 17, 2026 at 9:50 PM
[速報]PostgreSQL互換のサーバレスでOLTPとOLAP両対応を実現する「Microsoft Fabric HorizonDB」パブリックプレビュー | Publickey
[速報]PostgreSQL互換のサーバレスでOLTPとOLAP両対応を実現する「Microsoft Fabric HorizonDB」パブリックプレビュー
マイクロソフトは日本時間11月18日未明に開幕した年次イベント「Microsoft Ignite 2026」で、PostgreSQL互換のサーバレスデータベース「Microsoft Fabric HorizonDB」のパブリックプレビュー版...
www.publickey1.jp
November 17, 2026 at 10:14 PM
> [速報]PostgreSQL互換のサーバレスでOLTPとOLAP両対応を実現する「Microsoft Fabric HorizonDB」パブリックプレビュー
https://www.publickey1.jp/blog/26/postgresqloltpolapmicrosoft_fabric_horizondb.html
[速報]PostgreSQL互換のサーバレスでOLTPとOLAP両対応を実現する「Microsoft Fabric HorizonDB」パブリックプレビュー
マイクロソフトは日本時間11月18日未明に開幕した年次イベント「Microsoft Ignite 2026」で、PostgreSQL互換のサーバレスデータベース「Microsoft Fabric HorizonDB」のパブリックプレビュー版...
www.publickey1.jp
November 18, 2026 at 3:40 AM
Claude Code on the web is pretty decent at running web apps in its Playwright environment, but sometimes that’s not enough; edge cases often only appear in a real browser.

So I had @muninn.austegard.com craft this purely client side logger, following the OLTP standard: austegard.com/web-utilitie...
Agent Log | Oskar Austegard
Record a web page's console output, errors and network calls as OpenTelemetry logs in your browser, then hand the export to an AI agent to debug the page.
austegard.com
November 15, 2026 at 1:46 AM
[速報]PostgreSQL互換のサーバレスでOLTPとOLAP両対応を実現する「Microsoft Fabric HorizonDB」パブリックプレビュー - Publickey

https://www.walknews.com/1429250/

マイクロソフトは日本時間11月18日未明に開幕した年次イベント「Microsoft Ignite 2026」で、PostgreSQL互換のサーバレスデータベース「Microsoft Fabric HorizonDB」のパ [...]
[速報]PostgreSQL互換のサーバレスでOLTPとOLAP両対応を実現する「Microsoft Fabric HorizonDB」パブリックプレビュー - Publickey - WALK NEWS
マイクロソフトは日本時間11月18日未明に開幕した年次イベント「Microsoft Ignite 2026」で、PostgreSQL互換のサーバレスデータベース「Microsoft Fabric HorizonDB」のパブリックプレビュー版を発表しました。
www.walknews.com
November 17, 2026 at 10:48 PM
There are more historical examples of OLTP→OLAP (Oracle, Actian, Sybase, Informix, Couchbase), but that's because the OLTP DBMS market is much older (+50yrs) than the OLAP DBMS market (+30yrs).
November 11, 2026 at 5:41 PM
AWS Graviton5 (M9g) instances for Amazon RDS PostgreSQL deliver 121% higher throughput, 55% lower latency, and 92% better price-performance compared to Graviton2 (M6g) in OLTP benchmarks.
Boosting Amazon RDS for PostgreSQL with AWS Graviton5: Benchmarks
AWS Graviton5 (M9g) instances for Amazon RDS PostgreSQL deliver 121% higher throughput, 55% lower latency, and 92% better price-performance compared to Graviton2 (M6g) in OLTP benchmarks.
aws-news.com
November 13, 2026 at 12:14 AM
Looks ok. Doesn't look like it supports Entra ID as a man-in-the-middle though. MS is pushing everyone to use Entra ID and not SQL logins. Not everyone would be happy with change tracking either. And for many OLTP systems, it would just add another hop for each query.
November 12, 2026 at 10:26 PM
Why your team keeps arguing about Phoenix vs ClickHouse: one engineer breaks down the real performance trade-offs beyond the OLTP/OLAP labels.

https://dev.to/alnovis/phoenix-vs-clickhouse-why-the-fight-never-ends-1ogf

#data #analytics
November 11, 2026 at 8:00 PM
"Why databases store their data differently": @jpountz.bsky.social gives a high level overview of the main approaches to storing data with the use cases they're most suited for. jpountz.github.io/2026/04/19/w...
Why databases store their data differently
I was recently chatting with a friend about database architecture, and why, for instance, OLTP databases and search engines are implemented very differently. This friend is familiar with computer scie...
jpountz.github.io
November 6, 2026 at 8:12 AM
November 3, 2026 at 8:20 PM
idgaf if you run OLAP workloads on postgres but please do not run them on my OLTP database

no not even the readers, unless you are Very Clever about MVCC
November 2, 2026 at 6:25 AM
Postgres is still the right decision for like 60% of OLAP setups as well until the data starts scaling up A LOT or you need more governence, but definitely not on the same prod OLTP. It's just so dollar efficient compared to OLAP

DuckDB is somewhat changing this answer though nowadays
November 2, 2026 at 5:47 PM
Spot parallel query worker saturation via pg_stat_activity, where eager workers starve OLTP throughput. Tune max_parallel_workers_per_gather and max_worker_processes. https://www.valtersit.com/vault/analyze-parallel-query-worker-saturation-with-pgstatactivity-4dc46a/
Analyze Parallel Query Worker Saturation with pg_stat_activity
www.valtersit.com
November 2, 2026 at 9:10 AM
Grafana dashboards comparing ML and Bayesian forecasts on the same bout, as well as general dashboarding to see current events (on OLTP), Analytics (on Warehouse-Data), and forecast comparisons

All of it runs in Docker with hard memory caps, next to Pi-hole, on 8 GB of RAM.
The themed 6/
November 1, 2026 at 5:45 PM
I know I am repeating myself but AI-powered coding is insane. I've long been wanting go set up a small toy project that uses Kafka for streaming, an OLTP and an OLAP DB for transactions and warehouse (both ended up postgres lol) plus ETL with DBT and some dashboarding on top.
Before AI, 1/
October 31, 2026 at 8:33 PM
Karen Jex shows how to tune Postgres analytics without making OLTP users wait for their coffee. Use ANALYZE wisely. Control connections and memory then consider replicas or materialized views.\nhttps://postgr.es/p/9zJ
Karen Jex: Tuning Postgres for Analytics - PGConf.EU
https://postgr.es/p/9zJ
postgr.es
October 31, 2026 at 2:03 PM
Snowflake calls it Streams, and in Databricks it’s Change Data Feed. Another one is Databricks Lakebase - which is different name for OLTP tables.
October 30, 2026 at 4:03 PM
the big drawback of oltp on s3 is you're emulating one type of database on top of another, and there's an impedance mismatch

but the mst is already paying that cost, so the blow is lessened. you can always cache the record keys and values in memory or on nvme
October 29, 2026 at 4:11 PM