![](https://cdn.prod.website-files.com/6555d1bced7999bb4a8b174f/65c4e430e7d206a0bdf7c3e9_AWS_re-Invent_Lambda.jpeg)
AWS re:Invent Recap #1
As selected by our AWS Community Hero, Ian Mckay, here are the top 5 highlights from the AWS re:Invent opening keynote by Andy Jassy.
- Surprisingly, per-ms billing will have less of an impact than the actual minimum billable time. Many Lambda functions are short and quick and will run well below 100ms. Having the minimum at 1ms creates an ecosystem where some bills will be sliced by some factor.
https://aws.amazon.com/blogs/aws/new-for-aws-lambda-1ms-billing-granularity-adds-cost-savings/
- The new gp3 EBS volumes allow for independent scaling of capacity and performance. I'm glad I don't have to explain the gp2 calculation to customers any longer, and getting a small price cut to baseline performance is a bonus too.
- Earlier this month, I was burnt by Docker Hubs new rate limits. Thankfully, ECR Public will fix these issues immediately. Well done to everyone who made this happen, I know you were given a somewhat unreasonable deadline, but still managed to deliver.
https://aws.amazon.com/blogs/aws/amazon-ecr-public-a-new-public-container-registry/
- Lambda's increase to 10GB of RAM and 6 vCPUs opens the door for some workloads that were just out of reach previously. Combined with EFS support, this makes Lambda that much more powerful. Remember, still scale horizontally if you can!
- Aurora Serverless V2 has to be, in my opinion, the most impactful thing I've seen in the keynote. This is scaling performance in a fraction of a second. Does this also allow for sleep-at-night workloads without fear from the first morning users latency? We’ll see!
TAGS
AWS
Open Source
Community
DISCOVER MORE