Episode 36 - Humio at Stibo Systems with Torben Haagh
August 13th, 2020
John visits with Torben Haagh, Section Manager of Cloud Platform & Data Science at Stibo Systems, the master data management company that helps companies create transparency in their business processes. Hear how Torben uses Humio to help keep their new cloud-native content syndication platform stable and secure for customers.
Stibo Systems was founded in 1794 as a printing company for the church and the university, and the subsidiary that Torben is in was founded in 1976 to support the printing of catalogs, phone books, and other data-rich publications. They are an enterprise software company, focusing on data and data management, specifically master data management, specializing in product data, information, customer data information, product life cycle management, and product data syndication. They have big customers from all over the world, including Amazon, to Walmart, to the Home Depot. If you go online and look up a product that you want to buy, there's a good chance that the data flowed through the Stibos System data management system.
Torben is working on two product tracks.
“We have the one product that is a pure cloud-native solution that is based on microservices. It can scale individually, to the needs of the customer. The on-premise solution is auto componentized, but using more traditional ways than microservices, and we are moving that to the cloud also. We are also working on how to have a proper SaaS experience around our core enterprise platform.”
He describes some of the important aspects of the system he’s helping build for their customers.
“The cloud-native solution we made for the product data syndication. It has a tenancy system that's very dynamic. That's why we are transitioning the on-prem solution to the cloud gradually, and then doing things a little different than when you built new solutions. … What we see on the sales side is a huge shift from people that want to run it on-prem themselves to people that want it as a software-as-a-service solution.”
Stibo Systems customers naturally want 24/7 support, and Humio helps provide it in the cloud-native system for a new syndication platform. They are looking at it for aggregating across the enterprise application as well. They have Humio running in production on only one node, and they have a few nodes in their test environments as well.
Torben looks back at the log management system they had installed prior to trying Humio.
“We were looking at the ELK stack in the beginning and had it installed, but most developers actually turned directly to console rocks instead of using ELK, because it was too cumbersome and too tedious.”
Since using Humio, they fully understand the benefits of using index-free technology, which allows them to search for anything in the data without heavy indexes or defining what to store upfront.
“Everyone turns to Humio to figure out what is going on. Its ability to brute force search makes it so that you don't have to enrich the data beforehand. We originally had that problem in our application, because we used Elasticsearch for searching in our application. We know the pain about needing to define what you can be searching for in the future.”
“So the ability to just create a field with a regex expression on the fly and to create a chart that looks at specific issues that way creates transparency. It gives a really good understanding of what is going on. In that way, it is fairly easy for us to get an overview of the communication in the microservice system.”
Torben comments on how easy it is for his team to use Humio.
“Humio is really essential. Just go in and do a query on error logs right. Do a timespan query, an attend ID. You might see it immediately. So, really, really, really easy, because it's so easy to zoom in on the problem from very few parameters."
“You know, we don't do any training in Humio at all. People simply pick it up, themselves. That's easier.”
Listen to the rest of the podcast to answer these questions:
How do they make sense of a problem when all they get is a heap dump?
How do you solve a murder mystery when the body keeps disappearing?
How can Stibos Systems use Humio to fix customer issues swiftly, before they experience them themselves or have to call in?
How can even a slick hotfix process disturb strategic work and become a big noisy squeaky wheel?
Why did a university professor invite Kresten Krab Thorup to show him its unique architecture, and how long did it take him to install it himself in his cloud test environment (hint, it’s minutes, not hours)?
Why is Torben such a fan of Nikki Watt, CTO at OpenCredo, and why is her YouTube video Evolving Your Infrastructure with Terraform, and her talk about the need for insights in a cloud-native architecture (Journeys to cloud native architecture) worth a look?
How do they approach observability, and how do they use Humio together with Prometheus and Jaeger to “keep their stuff together, to keep from getting into big problems?"
Why is serverless technology so impactful, and how is event-driven thinking a huge mental leap, and changing the way software is developed?
What is the best way to stay close to what’s happening in the cloud-native community, and why is it worth taking time to get involved with groups like theCloud Native Aarhus Meetup?