An Exploration Of Tobias' Experience In Building A Data Lakehouse From Scratch

Summary Five years of hosting the Data Engineering Podcast has provided Tobias Macey with a wealth of insight into the work of building and operating data systems at a variety of scales and for myriad purposes. In order to condense that acquired knowledge into a format that is useful to everyone Scott Hirleman turns the tables in this episode and asks Tobias about the tactical and strategic aspects of his experiences applying those lessons to the work of building a data platform from scratch. Announcements Hello and welcome to the Data Engineering Podcast, the show about modern data management When you're ready to build your next pipeline, or want to test out the projects you hear about on the show, you'll need somewhere to deploy it, so check out our friends at Linode. With their new managed database service you can launch a production ready MySQL, Postgres, or MongoDB cluster in minutes, with automated backups, 40 Gbps connections from your application hosts, and high throughput SSDs. Go to dataengineeringpodcast.com/linode today and get a $100 credit to launch a database, create a Kubernetes cluster, or take advantage of all of their other services. And don't forget to thank them for their continued support of this show! Atlan is the metadata hub for your data ecosystem. Instead of locking your metadata into a new silo, unleash its transformative potential with Atlan's active metadata capabilities. Push information about data freshness and quality to your business intelligence, automatically scale up and down your warehouse based on usage patterns, and let the bots answer those questions in Slack so that the humans can focus on delivering real value. Go to dataengineeringpodcast.com/atlan today to learn more about how Atlan’s active metadata platform is helping pioneering data teams like Postman, Plaid, WeWork & Unilever achieve extraordinary things with metadata and escape the chaos. Struggling with broken pipelines? Stale dashboards? Missing data? If this resonates with you, you’re not alone. Data engineers struggling with unreliable data need look no further than Monte Carlo, the leading end-to-end Data Observability Platform! Trusted by the data teams at Fox, JetBlue, and PagerDuty, Monte Carlo solves the costly problem of broken data pipelines. Monte Carlo monitors and alerts for data issues across your data warehouses, data lakes, dbt models, Airflow jobs, and business intelligence tools, reducing time to detection and resolution from weeks to just minutes. Monte Carlo also gives you a holistic picture of data health with automatic, end-to-end lineage from ingestion to the BI layer directly out of the box. Start trusting your data with Monte Carlo today! Visit dataengineeringpodcast.com/montecarlo to learn more. Your host is Tobias Macey and today I'm being interviewed by Scott Hirleman about my work on the podcasts and my experience building a data platform Interview Introduction How did you get involved in the area of data management? Data platform building journey Why are you building, who are the users/use cases How to focus on doing what matters over cool tools How to build a good UX Anything surprising or did you discover anything you didn't expect at the start How to build so it's modular and can be improved in the future General build vs buy and vendor selection process Obviously have a good BS detector - how can others build theirs So many tools, where do you start - capability need, vendor suite offering, etc. Anything surprising in doing much of this at once How do you think about TCO in build versus buy Any advice Guest call out Be brave, believe you are good enough to be on the show Look at past episodes and don't pitch the same as what's been on recently And vendors, be smart, work with your customers to come up with a good pitch for them as guests... Tobias' advice and learnings from building out a data platform: Advice: when considering a tool, start from what are you actually trying to do. Yes, everyone has tools they want to use because they are cool (or some resume-driven development). Once you have a potential tool, is the capabilty you want to use a unloved feature or a main part of the product. If it's a feature, will they give it the care and attention it needs? Advice: lean heavily on open source. You can fix things yourself and better direct the community's work than just filing a ticket and hoping with a vendor. Learning: there is likely going to be some painful pieces missing, especially around metadata, as you build out your platform. Advice: build in a modular way and think of what is my escape hatch? Yes, you have to lock yourself in a bit but build with the possibility of a vendor or a tool going away - whether that is your choice (e.g. too expensive) or it literally disappears (anyone remember FoundationDB?). Learning: be prepared for tools to connect with each other but the connection to not be as robust as you want. Again, be prepared to have metadata challenges especially. Advice: build your foundation to be strong. This will limit pain as things evolve and change. You can't build a large building on a bad foundation - or at least it's a BAD idea... Advice: spend the time to work with your data consumers to figure out what questions they want to answer. Then abstract that to build to general challenges instead of point solutions. Learning: it's easy to put data in S3 but it can be painfully difficult to query it. There's a missing piece as to how to store it for easy querying, not just the metadata issues. Advice: it's okay to pay a vendor to lessen pain. But becoming wholly reliant on them can put you in a bad spot. Advice: look to create paved path / easy path approaches. If someone wants to follow the preset path, it's easy for them. If they want to go their own way, more power to them, but not the data platform team's problem if it isn't working well. Learning: there will be places you didn't expect to bend - again, that metadata layer for Tobias - to get things done sooner. It's okay to not have the end platform built at launch, move forward and get something going. Advice: "one of the perennial problems in technlogy is the bias towards speed and action without necessarily understanding the destination." Really consider the path and if you are creating a scalable and maintainable solution instead of pushing for speed to deliver something. Advice: consider building a buffer layer between upstream sources so if there are changes, it doesn't automatically break things downstream. Tobias' data platform components: data lakehouse paradigm, Airbyte for data integration (chosen over Meltano), Trino/Starburst Galaxy for distributed querying, AWS S3 for the storage layer, AWS Glue for very basic metadata cataloguing, Dagster as the crucial orchestration layer, dbt Contact Info LinkedIn Parting Question From your perspective, what is the biggest gap in the tooling or technology for data management today? Closing Announcements Thank you for listening! Don't forget to check out our other shows. Podcast.__init__ covers the Python language, its community, and the innovative ways it is being used. The Machine Learning Podcast helps you go from idea to production with machine learning. Visit the site to subscribe to the show, sign up for the mailing list, and read the show notes. If you've learned something or tried out a project from the show then tell us about it! Email [email protected]) with your story. To help other people find the show please leave a review on Apple Podcasts and tell your friends and co-workers Links Data Mesh Community Podcast OSI Model Schemata Podcast Episode Atlan Podcast Episode OpenMetadata Podcast Episode Chris Riccomini The intro and outro music is from The Hug by The Freak Fandango Orchestra / CC BY-SASponsored By:MonteCarlo:  Struggling with broken pipelines? Stale dashboards? Missing data? If this resonates with you, you’re not alone. Data engineers struggling with unreliable data need look no further than Monte Carlo, the leading end-to-end Data Observability Platform! Trusted by the data teams at Fox, JetBlue, and PagerDuty, Monte Carlo solves the costly problem of broken data pipelines. Monte Carlo monitors and alerts for data issues across your data warehouses, data lakes, dbt models, Airflow jobs, and business intelligence tools, reducing time to detection and resolution from weeks to just minutes. Monte Carlo also gives you a holistic picture of data health with automatic, end-to-end lineage from ingestion to the BI layer directly out of the box. Start trusting your data with Monte Carlo today! Visit [dataengineeringpodcast.com/montecarlo](https://www.dataengineeringpodcast.com/montecarlo) to learn more.Atlan:  Have you ever woken up to a crisis because a number on a dashboard is broken and no one knows why? Or sent out frustrating slack messages trying to find the right data set? Or tried to understand what a column name means? Our friends at Atlan started out as a data team themselves and faced all this collaboration chaos themselves, and started building Atlan as an internal tool for themselves. Atlan is a collaborative workspace for data-driven teams, like Github for engineering or Figma for design teams. By acting as a virtual hub for data assets ranging from tables and dashboards to SQL snippets & code, Atlan enables teams to create a single source of truth for all their data assets, and collaborate across the modern data stack through deep integrations with tools like Snowflake, Slack, Looker and more. Go to [dataengineeringpodcast.com/atlan](https://www.dataengineeringpodcast.com/atlan) and sign up for a free trial. If you’re a data engineering podcast listener, you get credits worth $3000 on an annual subscription.Linode:  Your data platform needs to be scalable, fault tolerant, and performant, which means that you need the same from your cloud provider. Linode has been powering production systems for over 17 years, and now they’ve launched a fully managed Kubernetes platform. With the combined power of the Kubernetes engine for flexible and scalable deployments, and features like dedicated CPU instances, GPU instances, and object storage you’ve got everything you need to build a bulletproof data pipeline. If you go to: [dataengineeringpodcast.com/linode](https://www.dataengineeringpodcast.com/linode) today you’ll even get a $100 credit to use on building your own cluster, or object storage, or reliable backups, or… And while you’re there don’t forget to thank them for being a long-time supporter of the Data Engineering Podcast!Support Data Engineering Podcast