<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/netflix-cuts-cassandra-read-latency-from-seconds-to-milliseconds-query-http-method-standardized-2r8sos2yr" -->

---
title: Netflix cuts Cassandra read latency from seconds to...
description: Netflix shipped a dynamic partition-splitting mechanism for Cassandra that dropped average read latency from seconds to low double-digit milliseconds on...
canonical: https://daily.dev/posts/netflix-cuts-cassandra-read-latency-from-seconds-to-milliseconds-query-http-method-standardized-2r8sos2yr
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Netflix cuts Cassandra read latency from seconds to milliseconds, QUERY HTTP method standardized | daily.dev
og:description: Netflix shipped a dynamic partition-splitting mechanism for Cassandra that dropped average read latency from seconds to low double-digit milliseconds on...
og:url: https://daily.dev/posts/netflix-cuts-cassandra-read-latency-from-seconds-to-milliseconds-query-http-method-standardized-2r8sos2yr
og:image: https://api.daily.dev/og/posts/2R8SOS2Yr.png
og:image:alt: Netflix cuts Cassandra read latency from seconds to milliseconds, QUERY HTTP method standardized
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.

# Netflix cuts Cassandra read latency from seconds to milliseconds, QUERY HTTP method standardized

**[Backend Digest](https://daily.dev/sources/backend_digest)** · 5 min read · 0 upvotes · 0 comments

## Summary

Netflix shipped a dynamic partition-splitting mechanism for Cassandra that dropped average read latency from seconds to low double-digit milliseconds on time-series workloads. RFC 10008 standardized the QUERY HTTP method in June 2026, the first new standard HTTP verb since PATCH in 2010. Debezium 3.6 is out with breaking changes across multiple connectors. AWS Certificate Manager now supports ACMEv2 for automated 45-day certificate issuance and renewal.

## Content

**TLDR:** Netflix shipped a dynamic partition-splitting mechanism for Cassandra that dropped average read latency from seconds to low double-digit milliseconds on time-series workloads. RFC 10008 standardized the QUERY HTTP method in June 2026, the first new standard HTTP verb since PATCH in 2010. Debezium 3.6 is out with breaking changes across multiple connectors. AWS Certificate Manager now supports ACMEv2 for automated 45-day certificate issuance and renewal.

---

## Netflix dynamic Cassandra partition splitting drops read latency from seconds to low double-digit milliseconds

Netflix built an automatic partition-splitting system for Apache Cassandra that detects oversized partitions and asynchronously splits them into smaller child partitions, with a metadata layer routing reads transparently — no application changes or downtime required. On their TimeSeries Abstraction platform, average read latency fell from seconds to low double-digit milliseconds, tail latencies dropped under 200ms, and CPU utilization and read timeouts both decreased. The initial implementation targets immutable partitions for safety, with mutable partition support and reprocessing of failed splits planned next. If you're running Cassandra for time-series workloads and hitting wide-partition pain, this is worth reading closely. [Read more](https://daily.dev/posts/foN4JRhm4)

## RFC 10008 standardizes the QUERY HTTP method, first new HTTP verb since PATCH in 2010

The QUERY method is now a standard HTTP verb as of June 2026, filling a gap that's caused real architectural awkwardness for years: read-only operations that need a request body. Unlike POST, QUERY is safe, idempotent, and cacheable — it never modifies server state and can be safely replayed. It also supports saving queries as reusable URIs. Practical adoption is expected around 2027-2028 as server and client implementations catch up, but it's worth factoring into API design decisions now. [Read more](https://daily.dev/feed-by-ids?id=DupkyDtLO&id=AQ9tUPYyH)

## Debezium 3.6 released with breaking changes across MySQL, Oracle, PostgreSQL, and MongoDB connectors

Debezium 3.6 is generally available with a broad set of new features and several breaking changes worth auditing before upgrading. MySQL drops the 'never' snapshot mode, Oracle deprecates the 'redo_log_catalog' mining strategy, PostgreSQL enum values now return in logical sort order, and MongoDB schema names shift for Avro compatibility. On the new features side: off-heap RocksDB-based schema history storage, quantile metrics across all connectors, a new Amazon SNS sink, JDBC sink as a native Debezium Server sink, an incubating Apache Fluss sink, a new YashanDB source connector, and OpenTelemetry support in the Debezium Platform. The Kafka-less pipeline path via JDBC sink in Debezium Server is also now documented with a full PostgreSQL-to-MySQL tutorial. [Read more](https://daily.dev/feed-by-ids?id=sME6JGEVp&id=5yi0fiMDI)

## AWS Certificate Manager adds ACMEv2 support for automated 45-day certificate issuance

ACM now supports the ACMEv2 protocol, letting you provision a managed ACME server endpoint that issues 45-day public TLS certificates from Amazon Trust Services. It works with Certbot, cert-manager for Kubernetes, and acme.sh out of the box. PKI administrators can define domain scopes, enforce wildcard policies, and delegate certificate requests to application teams without sharing DNS credentials — domain validation happens once at the endpoint level. Available in all commercial AWS regions now. [Read more](https://daily.dev/posts/kEhJ4cONk)

---

## Also notable

- **Elastic agentic SOC: specialized agents cost 5.7x less than skill-delegated agents for automated alert triage:** Elastic's production data from 36,822 real SOC conversations shows specialized agents cost $0.69 vs $3.42 per Windows endpoint investigation, and traced process ancestry 100% of the time vs 50% for the skill-delegated approach — the cost gap comes from LLM call count (4 vs 12 calls), driven by on-demand skill loading triggering extra deliberation steps that each pay the full growing conversation-history cost. [Read more](https://daily.dev/posts/PIx1hweMU)
- **MCP Enterprise-Managed Authorisation promoted to stable, Okta is first supported identity provider:** The MCP team's EMA extension is now stable, replacing per-server OAuth consent prompts with a centralised zero-touch flow using ID-JAG token exchange — but EMA only controls connection-level access, not per-action runtime authorisation, so separate controls are still required for what agents do once connected. [Read more](https://daily.dev/posts/nZMGaNvIP)
- **Cloudflare Workers Cache launches with tiered caching and stale-while-revalidate, no additional cost:** Workers Cache sits directly in front of Worker entrypoints, uses standard Cache-Control and Cache-Tag headers, supports automatic tiered caching across Cloudflare's global network, and cache hits skip Worker execution entirely — saving CPU billing — available now on all plans at no additional cost beyond standard request rates. [Read more](https://daily.dev/posts/iuM8TvTYb)
- **Microsoft, Google, and Cloudflare move quantum-safe cryptography deadlines forward to 2029:** All three have accelerated their post-quantum cryptography timelines citing faster-than-expected quantum hardware advances and harvest-now-decrypt-later attacks; Google is already integrating ML-DSA signatures in Android 17, and most organizations still lack visibility into where cryptography exists across their systems. [Read more](https://daily.dev/posts/R6dMKEgc7)
- **Databricks agent triage fleet saves 6,500 analyst hours in 30 days, agent-escalated alerts 10x more likely to be true positives:** Databricks' internal security team built 17 source-specific agents triaging 100% of low-severity alerts in real time using Spark Structured Streaming and deterministic pre-filtering before any LLM call, with agent-escalated low-severity alerts proving 10x more likely to be true positives than existing HIGH/MEDIUM alerts. [Read more](https://daily.dev/posts/xGK54FnSN)

## Similar posts on daily.dev

- [Netflix Cuts Cassandra Read Latency from Seconds to Milliseconds with Dynamic Partition Splitting](https://daily.dev/posts/netflix-cuts-cassandra-read-latency-from-seconds-to-milliseconds-with-dynamic-partition-splitting-fon4jrhm4) · InfoQ · 10 upvotes · 0 comments

---

Tags: [#aws](https://daily.dev/tags/aws), [#apache-cassandra](https://daily.dev/tags/apache-cassandra), [#debezium](https://daily.dev/tags/debezium), [#change-data-capture](https://daily.dev/tags/change-data-capture)

[View this post on daily.dev](https://daily.dev/posts/netflix-cuts-cassandra-read-latency-from-seconds-to-milliseconds-query-http-method-standardized-2r8sos2yr)

```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":"DiscussionForumPosting","mainEntityOfPage":"https://daily.dev/posts/netflix-cuts-cassandra-read-latency-from-seconds-to-milliseconds-query-http-method-standardized-2r8sos2yr","headline":"Netflix cuts Cassandra read latency from seconds to milliseconds, QUERY HTTP method standardized","text":"Netflix shipped a dynamic partition-splitting mechanism for Cassandra that dropped average read latency from seconds to low double-digit milliseconds on time-series workloads. RFC 10008 standardized the QUERY HTTP method in June 2026, the first new standard HTTP verb since PATCH in 2010. Debezium 3.6 is out with breaking changes across multiple connectors. AWS Certificate Manager now supports ACMEv2 for automated 45-day certificate issuance and renewal.","url":"https://daily.dev/posts/netflix-cuts-cassandra-read-latency-from-seconds-to-milliseconds-query-http-method-standardized-2r8sos2yr","datePublished":"2026-07-07T04:17:59.299Z","dateModified":"2026-07-07T04:18:19.233Z","author":{"@type":"Organization","name":"Backend Digest","logo":"https://media.daily.dev/image/upload/s--9uCSrsAb--/f_auto,q_auto/v1773839398/logos/backend_digest?_a=BAMAMiiu0","url":"https://daily.dev/sources/backend_digest"},"interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"isPartOf":{"@type":"WebPage","url":"https://daily.dev/sources/backend_digest","name":"Backend Digest"}}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Backend Digest","item":"https://daily.dev/sources/backend_digest"},{"@type":"ListItem","position":3,"name":"Netflix cuts Cassandra read latency from seconds to milliseconds, QUERY HTTP method standardized"}]}
```

