A First Look at Aurora Serverless v2

During the latest re:Invent, AWS announced the preview of next version of Amazon Aurora Serverless that scales in fraction of a second and will introduce multi-AZ support, global databases, and read replicas to the serverless world.

In my talk at Percona Live 2021, today I tried to answer the following questions:

  • What are the differences between v1 and v2?
  • Why did AWS introduced an entirely new product?
  • Is Aurora Serverless v1 a service to be already forgotten?
  • A simple test on Serverless v1 versus Serverless v2: any surprise?

Below are my slides, video will be available online soon. Thanks to all the ones who attended the talk!

Starting to play with Aurora Serverless v2

One of the key system variables you have to deal with MySQL is innodb_buffer_pool_size, the size of the buffer pool, the memory area where InnoDB caches table and index data. You try to learn how to set a sensible value and you usually hope you do not need to change it too often.

One day you start to play with serverless databases and everything changes: you can see the value of innodb_buffer_pool_size changing real time, every few seconds. According to the load of your database cluster.

Last week I discussed at DevOpsCon London some of the advantages and disadvantages of running managed MySQL services, including the preview of Aurora Serverless v2.

Next month I will dig deeper at Percona Live: looking at a direct correlation in the metrics between the capacity, the value of innodb_buffer_pool_size and the load of the database is just an interesting start.

Here are the slides of my talk at DevOpsCon and my InfoQ news article about the preview of Aurora Serverless v2.

InfoQ – April 2021

From Cloudflare new WAF to Grafana going AGPLv3, from Netflix
open-sourced ConsoleMe to OpenBuilt: a recap of the topics I covered for InfoQ in April 2021.

Amazon Elasticsearch Introduces Auto-Tune

Amazon has recently announced the Auto-Tune feature in Amazon Elasticsearch Service, a closed-loop control system that adapts the Elasticsearch cluster to the running workload. The new automated memory management provides better ingestion throughput for log analytics workloads and reduced tail latencies for search queries.

Cloudflare Announces New Web Application Firewall

Cloudflare has recently introduced a new Web Application Firewall. The latest engine is written in Rust, provides better performances and integrates with other Cloudflare products.

AWS Introduces EC2 Serial Console: Troubleshoot Boot and Networking Issues

AWS has recently introduced the EC2 Serial Console, a tool to establish a serial connection to EC2 instances and troubleshoot boot and network connectivity issues. The new feature is designed to help system and network administrators to address production issues.

Netflix Open Sources ConsoleMe to Manage Permissions and Access on AWS

Netflix has recently open-sourced ConsoleMe, a AWS multi-account management service, and its CLI utility, Weep. The tools provide a central control plane for permissions management across all of AWS accounts of an organization and help to implement the principle of least privilege.

IBM, Red Hat and Cobuilder Develop OpenBuilt, a Platform for the Construction Industry

IBM, Red Hat and Cobuilder recently announced a collaboration to develop OpenBuilt, a platform for the construction industry supply chains built on the hybrid cloud platform Red Hat OpenShift and running on IBM Cloud.

Grafana Labs Changes Licenses to AGPLv3 for Grafana, Loki, and Tempo

Grafana Labs has recently announced the plan to change the licenses for their core products. They will relicense Grafana, Grafana Loki, and Grafana Tempo from the Apache License 2.0 to the Affero General Public License (AGPL) v3. Plugins, agents, and certain libraries will remain Apache-licensed.

Serverless Architecture Con – The Hague 2021

A Journey to Serverless Databases.

Today I was (virtually) on stage at the Serverless Architecture Con – The Hague 2021, discussing the benefits and the drawbacks of running a serverless MySQL on AWS.

Serverless databases are an interesting option for lazy cloud architects and devops engineers. And being lazy on the cloud is (often) a good thing.

The slides of my talk are below.