#715: AWS News: Be your own data analyst with Amazon Q in Quicksight, and more

0:00This is episode 715 of the AWS podcast, released on April 7th, 2025. Hello, everyone, and welcome back to the Adabys Podcast. I'm going to share with you. Great to have you back. Joined by one of my illustrious co-hosts, Gillian Ford. G'day, Gillian. Illustrious co-host. I love it, Simon. Always excited to be here. That's good to have you here. We're operating in 2AZ mode at the moment. Shruti is not with us today, but she'll be back. Lots of new stuff to talk about today. The first one that we wanted to call out is a new capability in Amazon Q in QuickSight. So this is scenario analysis. Now, I'm a data person, but I often don't know what questions to ask.

0:42And looking at this one, Julian, this helps you figure out how to get the answers to questions that you don't have enough manipulative knowledge to figure out yourself. Yeah, I love that. Especially like, it's so easy if you start with like a spreadsheet and then to feel a bit overwhelmed and kind of limited by what a simple spreadsheet can do. And I love that Q and Q and Q. site can help you to be able to ask questions like, what if we extended our free trial period? What if we change suppliers? A lot of these things that are a lot harder to just do in a spreadsheet. It's true. It's like there's different factors.

1:17So, you know, one of the examples that we talk about is, you know, how might we increase our marketing ROI by 15 % or more? So that's kind of like beyond goal seeking. It's kind of like important stuff or how to reduce customer mature by 20%. So again, natural language. And then what Q will do inside QuickSight is it'll find the relevant data, it'll suggest approaches, it'll plan each step and execute, and it'll give you detailed insights along the way. So this ties into, I guess, agentic AI that we've been talking a lot about, about the fact that breaking down tasks into littler tasks and then having them happen automatically gives you a lot of power.

1:53And it just means you can get answers quicker, can't you? Yeah, or if you had to get these answers before, it would require you to maybe work with like a data scientist or maybe had to work with like a DBA, you could go and like run the queries for you. And now within QuickQ and QuickSight, as long as you have access to your QuickSight account, you can start and actually go and ask those questions. Exactly. And it creates an analysis canvas. So you can sort of go down different paths because finding answers is often and exploratory thing. You can read from QuickSight dashboards connected to Amazon Athena, Aurora, Redshift, the Simple Storage Service, OpenSearch service.

2:33It supports CSVs and single table Excel SX file uploads, so the old Excel spreadsheets. It's about 10 times faster than doing it the old-fashioned way, which is kind of nice. And you can also reuse, which I think is interesting if you're sort of discussing and comparing. You can go back and try things again. Pretty exciting stuff, Simon, I think. Yeah, it's nice. Nice stuff. Well, speaking of analytics, Jillian, why don't you kick us off with analytics? Amazon DataZone now supports metadata rules for publishing. Amazon OpenSearch service introduces two new instances, OR2 and OM2, expanding the OpenSearch Optimized Instance family.

3:12The new generation of OR2 instance delivers up to 26 % higher indexing throughput compared to previous OR1 instances and 70 % over R7G instances. The new OM2 instance delivers up to 15 % higher indexing throughput compared to OR1 instances and 66 % over M7G instances in internal benchmarks. AWS announces the general availability of PySpark in AWS cleanrooms. Let's talk about application integration. Amazon Coretto 24 is now generally available. This, as a reminder, is a no-cost multi-platform production-ready distribution of OpenJDK. This is a feature release, so it'll be supported through October 2025.

3:58We also have the long-term releases as well. This has a bunch of new updates. Speaking of new updates, artificial intelligence. There's a lot happening there. Amazon Bedrock Guard Rails has announced the general availability of industry-leading image content filters. So this allows you to block up to 88 % of harmful multimodal content. This new capability really removes that undifferentiated heavy lifting of building your own safeguards. And it also reduces the cycles you need to spend on manual content moderation, which can be both error-prone and tedious. So it'll allow you to do this automatically.

4:35You can deny and disallow specific topics. You can redact PII. You can block specific words. There's also contextual grounding checks to detect and block model hallucinations. There's a whole lot of stuff going on there. And you can explain the factual claims in model responses using the automated reasoning checks, which is very, very cool. Amazon SageMaker has introduced metadata rules to enforce standards and improve governance. So this allows you to understand what is going on with your data. By standardizing metadata practices, organizations can improve compliance, enhance your audit readiness, and streamline your workflows.

5:13Basically, data domain owners can define mandatory metadata fields that data users must complete when publishing assets to the catalog or requesting access to the data. Adibus Marketplace has introduced new seller experiences for machine learning products. Sellers can now quickly publish and update their ML listings with a new self-service experience, and they can also now create, view, and manage private offers for their ML products as well. Amazon Bedrock Knowledge Bases now supports Amazon OpenSearch Managed Cluster for vector storage. And Amazon Bedrock Custom Model Import has introduced real-time cost transparency, so now you can understand the compute resources being used to calculate inference costs in real time, which is pretty cool.

5:55There's a new Korean voice for Amazon Polly. I'm probably going to get the name terribly wrong. GHA is what it looks like to me. It's a neural text-to-speech female voice for Amazon Polly. We're happy to announce multi-head node support in Slurm for Amazon SageMaker hyperpod clusters. Couldn't be more jargon in that if I tried, I think. Basically, this new capability enhances fault tolerance and availability for large-scale Gen AI model development workloads. When a single head node manages job scheduling and resource allocation, it can become a bottleneck. And if the node is unresponsive or goes down, we've got lots of trouble.

6:32Now you can have one primary head, a controller, node to control all the others, and you can have additional backup head nodes on standby. If the primary fails, it'll switch across as well. The Research and Engineering Studio and AWS version 2503 is now available. So this gives you even more insight into any work you're doing from a research standpoint. point. Amazon Bedrock now supports RAG evaluation in GA. So you can evaluate your retrieval augmented generation applications, either built on Amazon Bedrock knowledge basis or a custom RAG system. And you can evaluate either retrieval or end-to-end generation.

7:10And these are powered by LLM as Judge with a choice of several judge models. So this allows you to understand how well you're doing Judge by another LLM. For retrieval, you can select from metrics like context relevance and coverage and for end-to-end retrieving generation you can select from quality metrics like correctness completeness hallucination detection and other responsible ar metrics like harmfulness answer refusal and stereotyping lots of capability in this one amazon bedrock model evaluation llm as judge is now generally available this one ties in with the previous one doesn't it so now you can evaluate compare and select the right models for use cases and you can use an LLM as your judge from several available to decide what is good or what is not so good.

7:56And Amazon Bedrock Guard Rails has announced policy-based enforcement for responsible AI. This is a new feature that enables customers to apply specific guard rails to model inference calls, ensuring responsible AI policies are applied across all AI interactions. This helps you protect against the nasties that come in, undesirable content, should you have topic filters, PIR detection, all that stuff. And guardrails can be applied across any foundation model, including those hosted with Amazon Bedrock, self-hosted models, and third-party models outside of Bedrock by using the Apply Guardrail API.

8:31So this is important because it gives you a consistent user experience and standardizes safety and privacy controls. Great. Let's talk about business applications. Amazon SES announces the availability of the Vade add-on for Mail Manager. This is a sophisticated content filter that enhances email security for both incoming and outgoing messages. This add-on, developed in collaboration with Hornet Security, combines heuristics, behavioral analysis, and machine learning to provide robust or protection against evolving communication threats such as spam, phishing attempts, and malware. Next up, we've got Compute.

9:09Amazon EC2 now supports up to the full EC2 instance bandwidth for inter-region VPC peering traffic and to AWS Direct Connect. Additionally, EC2 supports jumbo frames up to 8 ,500 bytes for cross-region VPC peering. This is a big deal for network geeks like myself. The ability to have jumbo frames means you can move more data quicker with less overhead, so I'm kind of excited about this one. Oh, I like it. Okay. Been a while since we've got some hype around a networking release. Really one so deep in the stack too, but hey, jumbo frames are a good thing. We're announcing Terraform support for AWS Parallel Computing Service.

9:51Customers can now use Terraform to create and manage their PCS clusters. AWS Deadline Cloud now supports IPv6. Let's talk a bit about contact sensors. There are three new features on the connected mobility solution on AWS. there is now role-based access control multi-account and multi-region support there is fleet management portaling preview as well so lots of things if you're into that sort of part of the world amazon connect contact lens now enables you to capture agent acknowledgements of performance evaluations so agents get an evaluation of their performance you know the old how did i do they can acknowledge the fact that they got it back and they can also add notes to it as well and adibus announces the next generation of Amazon Connect where powerful AI improves every interaction.

10:40Basically, every customer touchpoint now gets the opportunity to have a better outcome. So this approach spans self-service, agent assistance, analytics, post-contact evaluation, and automated follow-ups to boost sales and delight customers. And essentially, this is enabled with a single click and it has unlimited use of AI capabilities. So you can just figure out what the best customer experiences rather than trying to budget. A quick update in the world of containers. Amazon EKS now enforces upgrade insights checks as part of cluster upgrades. This is a significant step forward, giving cluster administrators confidence with your upgrade.

11:15You know, you're always a bit nervous when upgrades happen, so this lets you be less nervous. Next up is Database. Amazon DynamoDB now support percentile statistics for the successful request latency Amazon CloudWatch metric. This percentile statistic enables you to understand the latency distribution of your successful request to DynamoDB, complementing the existing average, minimum, and maximum statistics. Amazon DynamoDB Streams now comes with AWS private link support, allowing you to invoke DynamoDB Streams APIs from within your VPC without traversing the public internet. Amazon RDS for SQL Server now supports linked servers to Teradata databases.

12:04Linked Server is a SQL Server feature that enables customers to read data and execute commands on remote databases servers outside of the SQL Server instance. So with this launch, customers can link their RDS for SQL Server instance to a Teradata database running on AWS or on-premises. Amazon Keyspaces for Apache Cassandra has expanded its multi-region replication capabilities, now enabling you to replicate your tables beyond the previous quota of six AWS regions to all available AWS regions. Amazon RDS for MySQL now supports MySQL Innovation Release 9.2 in the Amazon RDS preview environment, allowing you to evaluate the latest innovation release on Amazon RDS for MySQL.

12:52Amazon Keyspaces for Apache Cassandra adds multi-region support for user-defined types. This enhancement allows you to add user-defined types consistently across multiple AWS regions, enabling global applications to maintain consistent data schemas. AWS Database Migration Service Schema Conversion is a fully managed feature of DMS that automatically assesses and converts database schemas to formats compatible with AWS target database services. Now we're excited to announce that schema conversion now supports conversions from IBM DB2 for Z-OS to Amazon RDS for DB2. Now, Julian, this is a big deal for old heads like me.

13:39Z-OS is the way we say it for us graybeards. but this is a big deal because moving from mainframe to open systems is not easy um you know i'm going to throw in words like big indian and little indian i'm also going to throw the word ebsidic in there as well which is probably going right over your head there but the formats of data on the mainframe are completely different to the open systems world and so this is this is a big deal so i'm really impressed with what the team got done here i am i mean i feel like i've been born in the cloud, I wouldn't even know where to get started. Yeah. You're lucky.

14:14Totally. Speaking of born in the cloud, Amazon DynamoDB Accelerator SDK for JavaScript version 3 is now available. And so is the SDK for Go version 2. Next up is developer tools. AWS CodeBuild now supports an enhanced S3 caching experience. you can now define custom cache keys for more granular cache management and improved cache persistence across your builds. AWS CloudFormation now supports targeted resource scans in the infrastructure-as-code generator. AWS Amplify Hosting announces web application firewall protection in general availability. Amplify Swift now supports sharing authentication state across multiple apps by leveraging keychain access groups.

15:06This new feature allows developers to manage a single authentication session across all Swift-based applications and extensions within the same access group. AWS announces increased quotas for AWS Client VPN, expanding routes per target network association to 100, and authorization rules per endpoint to 200. And one quick update in front-end web and mobile, AWS announces expanded service support in the AWS Console mobile app. Now with the mobile app, you can access 24 additional AWS services, including service quotas, CloudFront, Amazon SES, AWS Cloud9, and AWS Batch via an integrated mobile web browser experience in the AWS console mobile app.

15:53Let's talk about games. Amazon GameLift servers have expanded instance support with next-generation EC2 instance families. They now support the Amazon EC2 fifth through eighth generation instances. So you get lots more price performance, efficiency, and flexibility for hosting based on different types of service. Let's talk management and governance. Amazon CloudWatch ROM now supports monitoring multiple domains with a single app monitor. So this allows you to monitor multiple top level domains and second level domains using a single app monitor. And you can monitor everything across the one place.

16:30So one dashboard can see everything. That's pretty cool. And CloudWatch RUM, or RUM as I like to call it, now supports JavaScript source maps for easier error debugging so you can understand what the script was doing. One quick update in media services. AWS Elemental Media Connect adds support for NDI, or network device interface, outputs from Media Connect flows. Let's talk networking and content delivery. Adibus Network Manager and Adibus Cloud WAN now supports Adibus Private Link and IPv6. The ongoing V6 march continues. Adibus Network firewall adds pass action rule alerts and JA4 filtering.

17:11So this is the ability to generate alerts on traffic that matches pass action rules and fingerprinting in firewall rules as well. So this gives you enhanced visibility into your network traffic without having to add an alert action rule before the pass action rule. So this helps you figure out anomalies that would otherwise cause issues without additional scrutiny. Amazon Route 53 Profiles has introduced dual stack support for the Route 53 Profiles API endpoints, which means you can now connect using IPv6, IPv4, or dual stack clients as well. Existing Route 53 profile endpoints supporting IPv4 will remain available for backwards compatibility.

17:51The Adibus Network Firewall has introduced a new flow management feature. So this enables you to identify and control active network flows. So there are two key new functions, flow capture, which allows point in time snapshots of active flows and flow flush, which enables selective termination of specific connections. So as a customer, you can now view and manage your active flows based on the criteria like source or destination IP addresses, ports, protocols, and you get lots more control over what's going on in the network. And a quick update around payment, the option to have currency selection has been added to payment profiles.

18:31So this capability builds upon the existing ability to set different payment methods per service provider. So now you have more control. For example, you can select to pay one AWS service provider in USD and another in euros. So you can do it up to however you like. But if you don't mean to create a profile, you can also just use the default payment preferences. And a quick update for quantum technologies, IronQ Fort Enterprise is now available on Amazon Bracket. So this is IronQ's latest 36 qubit Fort Enterprise quantum processing unit. This device joins the existing quantum hardware profile on Bracket, which includes Fort1, ARIA1 and ARIA2.

19:12So again, you can use the familiar Bracket SDK and APIs to access this new device. This device features INQ's de-biasing and sharpening error mitigation algorithms to enable advanced customer workloads. And Fort Enterprise continues to use the native ZZ gate architecture, make it easy for customers to seamlessly migrate workloads between Fort devices. Look at this, it's quantum computing, and we're still doing migrations between devices. The Enterprise device is physically located in Switzerland, but all customer traffic routes through U.S. East, North Virginia. Next topic, security, identity, and compliance.

19:52AWS IAM announces a new dual-stack public endpoint, enabling customers to connect to IAM over the public internet using IPv6, IPv4, or dual-stack clients. AWS WAF now supports URI fragment field matching, enabling customers to match against the URI fragment and along with the already supported URI path. With this feature, customers can create rules that inspect and match against the content of the URI fragment within their URI path. Now on to serverless. AWS Lambda now supports creating serverless applications using Ruby 3.4. AWS Fault Injection Service now supports a recovery action for Amazon Application Recovery Controller Zonal Autoshift.

20:40A recovery action is a new fault injection service action type that allows customers to demonstrate how AWS responds during an availability incident. This sounds super useful. So an example, when AWS detects potential infrastructure issues in an availability zone, such as power or network disruptions, zonal auto shift automatically shifts traffic away from the availability zone. With the new Fault Injection Service Recovery action, customers that have enabled Zonal Autoshift can now run the Fault Injection Service of AZ Availability, the power interruption scenario, to induce the expected symptoms of a complete interruption of power in an availability zone and demonstrate how AWS would trigger Zonal Autoshift.

21:29This is so cool. This is going to allow customers to, yeah, tune their monitoring recovery process to improve resiliency and really be able to ensure based on their own SLAs that they're able to meet them. Amazon EventBridge Scheduler now supports AWS PrivateLink, providing you access to Scheduler from within your VPC without using the public Internet. it. Next up is storage. You can now use Amazon EBS GP3 and IO1 volumes in an AWS dedicated local zone. Dedicated local zones are a type of AWS infrastructure that are fully managed by AWS, built for exclusive use by you or your community and placed in a location or data center specified by you to help you comply with regulatory requirements.

22:21You can now easily deploy pre-configured sample applications to connect your users to data in Amazon S3 using Storage Browser for S3. Once you select and deploy one of the sample apps, users can browse, download, upload, copy, delete data that you have access to in S3 through an intuitive file browser experience. I am so happy that exists. Yeah. Yeah. And wow. That should have existed a long time ago. I don't know why it took us so long, but we got there. So I'm excited. So there are lots of updates today. Any that leapt out at you as something of like, oh, I'm going to get my hands on that one.

22:56It sounded like you kind of liked the AZ testing one. Oh, absolutely. I mean, I think there's something in this one for everyone, and that's where I get really excited. And especially I work with startups. I know a lot of startups that have goals of being able to scale, but they haven't really tested their resiliency strategy. So I love that this is something that they can be able to implement to really be able to ensure that they can meet their SLAs for their customers. What about you? That's nice. Well, for me, it was actually the AWS Amplify hosting with the WAF support going into general availability.

23:33Now, it's not an exciting thing in terms of, wow, this is an amazing new functionality. But for me, it's just make it easier to deploy something and be secure. I like that. So I like the easy button. The more easy button we can get, the better. Jillian, how do folks reach out to you? Jillian Ford on LinkedIn. That's the way. And also if you want to do it old school, AWS podcast at amazon.com. If you want to send us a question for our mailbag episode as well, please feel free to do that to the same address. And until next time, keep on building.