# Amazon Aurora

23 Hacker News commenters recommended Amazon Aurora and 31 criticized it, November 2025 to November 2026. It was recommended in 8 of 13 months.

**Verdict:** Managed relational database that handles storage, replication and failover, so teams skip most admin work. Critics say costs run high and are hard to predict.

## Why people recommend it

- Little operational work for teams (6 comments): "You lean heavily on managed services (Aurora Serverless, Kinesis, Step Functions) where the operational load is the value prop." ([source](https://news.ycombinator.com/item?id=45895558)) "Running all your postgres on ec2 is like 1000X harder than looking at the dashboard for an Aurora cluster." ([source](https://news.ycombinator.com/item?id=46529811)) "The experience of Aurora is I can have all of that , in like 30 lines of terraform and a few minutes." ([source](https://news.ycombinator.com/item?id=46900399))
- Storage layer handles replication and durability (4 comments): "bulk data loads are available on replicas a lot sooner on Aurora because the storage is the replication system, not the WAL" ([source](https://news.ycombinator.com/item?id=46338091)) "you are comparing systems that are already redundant and highly available (e.g. Aurora, writing six copies of the data)" ([source](https://news.ycombinator.com/item?id=48613735)) "Their storage layer worked perfectly and prevented the ACID violations." ([source](https://news.ycombinator.com/item?id=45933239))
- Robust and reliable in production (3 comments): "we've been running multiple aurora clusters in production for years now and have not encountered this issue with failovers" ([source](https://news.ycombinator.com/item?id=45930923)) "genuinely one of the most robust solutions out there in that space" ([source](https://news.ycombinator.com/item?id=46899308)) "It has been very robust for us." ([source](https://news.ycombinator.com/item?id=45931142))
- Fast on write-heavy workloads (3 comments): "Aurora is faster than MySQL on equivalent hardware, especially for write-heavy workloads, just that the “5× faster” claim only holds under very specific benchmark conditions." ([source](https://news.ycombinator.com/item?id=45995827)) "delivers significantly higher performance for the same underlying hardware" ([source](https://news.ycombinator.com/item?id=45992574)) "Aurora is fast enough that we actual started to contend on the heap locks" ([source](https://news.ycombinator.com/item?id=46527545))

## What people warn about

- Expensive, hard-to-predict I/O costs (5 comments): "What if we made people pay for storage and I/O? And we made estimating I/O practically impossible?" ([source](https://news.ycombinator.com/item?id=45980104)) "We were burned by Aurora. Costs, performance, latency, all were poor and affected our product." ([source](https://news.ycombinator.com/item?id=45931449)) "More often than not there's a single pg aurora instance responsible for 50%+ of the bill." ([source](https://news.ycombinator.com/item?id=49370786))
- Failover behaves unexpectedly at times (3 comments): "My cluster recently started to failover every few days whenever it experiences the load to trigger scale up" ([source](https://news.ycombinator.com/item?id=45933962)) "What's particularly concerning is that this seems to be a fundamental flaw in Aurora's failover mechanism" ([source](https://news.ycombinator.com/item?id=45930489)) "indicates something wrong with how they propagate DNS changes during the failover" ([source](https://news.ycombinator.com/item?id=45930256))
- Docs and behavior often surprise users (2 comments): "its not the first time I have noticed unexpected and odd behaviors from Aurora PostgreSQL offering" ([source](https://news.ycombinator.com/item?id=45937687)) "Their docs are not good and things frequently don't behave how you expect them to." ([source](https://news.ycombinator.com/item?id=45930865))
- Benchmarks overstate speed versus bare metal (3 comments): "Aurora is faster than MySQL on equivalent hardware, especially for write-heavy workloads, just that the “5× faster” claim only holds under very specific benchmark conditions." ([source](https://news.ycombinator.com/item?id=45995827)) "I don't think that these numbers stand if you compare with a PostgreSQL on a bare metal server with NVMe SSDs attached to it." ([source](https://news.ycombinator.com/item?id=46528445)) "you are comparing systems that are already redundant and highly available (e.g. Aurora, writing six copies of the data)" ([source](https://news.ycombinator.com/item?id=48613735))

## Alternatives hackers recommend

- [Amazon RDS](https://hackerapproved.com/p/amazon-rds.md): 50 recommend
- [Neon Postgres](https://hackerapproved.com/p/neon.md): 48 recommend
- [DynamoDB](https://hackerapproved.com/p/amazon-dynamodb.md): 46 recommend
- [Cloudflare D1](https://hackerapproved.com/p/cloudflare-d1.md): 36 recommend
- [PlanetScale](https://hackerapproved.com/p/planetscale.md): 33 recommend

## Where to get it

- [Visit the official site](https://aws.amazon.com/rds/aurora/)

## Cite as

Hacker Approved, "Amazon Aurora", https://hackerapproved.com/p/amazon-aurora/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).