A few short years ago, contactless car rentals were the realm of science fiction. If you wanted a car, you had to fill out and sign half a dozen forms and wait while an agent approved your credit card before you were handed the keys. Even if you booked a car online, you still needed to talk to a person and verify your personal information before you could get behind the wheel. These days, you can rent a car on a mobile app and then use your phone to unlock the vehicle, which is faster and, these days, safer.
Huf Secure Mobile GmbH (Liberkee, effective from February 2021) used to be a wholly owned subsidiary of Huf Hülsbeck & Fürst GmbH & Co. KG. Founded in 1908 in Velbert/North Rhine-Westphalia, the group of companies develops mechanical and electronic locking and tire pressure monitoring systems for the global automotive industry. Huf was the first company ever to design components for the vehicle access and driving authorization systems “Passive Entry” and “Keyless Go”. Huf is one of the 20 most innovative companies in Germany.
Our customers include car rental and sharing service providers and other operators of vehicle fleets. Our applications streamline the management of automotive fleets, provide usage metrics, and allow users keyless access using mobile phone apps while also securing vehicles against unauthorized use.
Learning Systems Engineering on the Job
I’ve been with Huf Secure Mobile (Liberkee) for just over four years now. It was my first job in a Systems Engineering role. Immediately, I found myself in a world of virtual machines, databases, and other unfamiliar IT concepts. I spent a lot of time teaching myself everything I could about that new subject, both on the job and after hours.
My diligence and enthusiasm paid off, and I was then promoted to Head of Digital Operations, rising to a position within the company that allowed me to set the tone for our operations activities and help the team transition to cloud technology.
Making Sense of an Ordered Mess
Our IT infrastructure has grown over time with changing business requirements to roughly 300 VMs running on bare metal on-prem servers with hundreds of services that we could not oversee and monitor in an appropriate way. Alerting ahead of customer complaints was a big issue. Several monitoring solutions were in place but were not fully integrated and fine-tuned. The whole system was working but required attention and ultimately could be streamlined to be operated by fewer people, thus lowering the operational costs.
We were dealing with physical and virtual servers that were partially documented and had little automation. For new employees joining the team, it was hard to get up to speed.
Formulating a Plan
We saw great potential in moving our application and system landscape towards containers and in October 2019, we brought the whole team together to map out our target architecture. The workshop led by Container Solutions helped us a lot to get moving very quickly.
It was then that we started talking about infrastructure-as-code and running some—if not all—of our services in a Kubernetes cluster. We had very experienced engineers on board as well as the server infrastructure, so we were able to set up our own Kubernetes cluster created via Terraform in a very short time frame and started to migrate our microservice-based applications into it.
Finally, we wanted to create an automated monitoring system that would trigger alerts when something went wrong on the back end, whether it was a piece of erroneous code, a service or an application shutting down, or a failed database call. We put together groups to discuss various solutions, and Application Performance Management was one of the desires that emerged. Application performance management lets you streamline your development cycle and rightsize your infrastructure.
Streamlining Development Cycles and Rightsizing Infrastructure Using APM
I had heard about application performance management (APM) before, but I didn’t know its ins and outs. Fortunately, our interim manager was well versed in this technology and had worked with it for over a decade. He showed us some of the tools he had used in the past, including New Relic and Datadog. But after reading its documentation, he recommended a more recent application performance management platform called Instana.
He liked the look and feel of Instana and recommended it as the fastest and easiest way to board the application performance management train. We reviewed the product documentation ourselves and watched some videos. Everything convinced us that Instana had all the features we needed in a single tool, including automated alerts (Smart Alerts) and the ability to monitor applications, services, and servers.
Up and Running in 60 Minutes
We ordered a trial and started deploying Instana agents on our virtual machines. It was so easy—all we had to do was copy and paste a one-line cURL command, and Instana automated the rest. Once installed, Instana natively supports and instantly detects dozens of database and application frameworks. Besides a couple of services that we had to add manually, Instana directly discovered most of the components that comprise our applications and our operations production environment.
Within an hour, we logged into the Instana dashboard and were able to view our first traces and metrics. We had gone from manually inspecting logs to our first taste of automated application performance monitoring in less than 60 minutes. Within a couple of days, we’d begun to create custom alerts based on the information extracted through Instana.
Following this successful trial, we expanded our use of Instana and installed it on our most critical systems, which are the ones that are most likely to impact our customers. If one of our client apps is crashing in the middle of the night, we got notified and could jump right into Instana, locate the issue, and take immediate action. This results in shorter outages and increased availability and functionality, allowing us to maintain SLAs for our customers, and improved products for their end-users.
Deep Visibility and Amazing Support
One of the most exciting developments is Instana’s Smart Alert feature. The general alerting of Instana was already super useful, but the Smart Alerts was what we really needed to efficiently create custom alerting on events in predefined Application Perspectives which can be used to group a specific set of services or requests.
The Release Marker is yet another of Instana’s awesome features that help you pinpoint the moment in time when a new release/deployment happened in the infrastructure.
What we found also really useful for people that don’t want to dive too deep into technical things is the “Custom Dashboard” feature. Instead of clicking through the entire Instana platform, they can create a customized view that displays only their traces and metrics they need to do their job.
Instana helps our developers work faster and smarter. We can set Release Markers on every new deployment, which can be used to quickly identify the moment in time of the deployment in the Instana dashboards, and see if the error rate or latencies increase caused by the newly deployed code. It also became a lot easier to follow automated and manual tests by having that insight in the traces.
We’ve been thrilled with the Instana platform, and also with their amazing customer support. We have monthly calls with our customer success manager, who provides valuable in-depth feedback on our uses. When we need immediate assistance, we can open a ticket on Instana’s support portal, and their team will get back to us in no time. There is no question too big or small for their experts to answer.
Instana is also very attentive when it comes to new feature requests. When we tell them what we need or what we’re missing, they work to integrate these functionalities into future versions of the platform. It’s great to have Instana’s ear because we can plan our next steps, secure in the knowledge that they’ll build the application performance monitoring tools we need to manage our operations environment and create new digital vehicle access products for our customers.
Actionable Insights Lead to Accelerated Deployment Cycles
The trend in the developer community towards DevOps clearly shows that there is much to learn and only picking up development skills is not enough. Instana harnesses automation, machine learning, and AI to sift through thousands of lines of code or page after page of logs—all for the sake of giving developers the visibility that they can’t otherwise access.
Cloud-based Apps are Too Complex to be Monitored Manually
Instana provides immediate, impactful, and actionable insights into how systems interact, thus alerting us to bottlenecks, conflicts, and coding errors that could bring our applications to a crashing halt.
Huf Secure Mobile’s (Liberkee’s) operations environment has evolved to the next level and everything is now easily visible for the business, and we have powerful tools to accelerate and improve the development and deployment cycle for all our applications. Instana has put our developers in the driver’s seat, and it helps us ship faster and more comfortably, with lower time-to-market stretches for new features. Therefore Instana has become a main pillar to support our customers on multiple continents.
This guest blog post was written by Jonas van Ravenstein, Head of Digital Operations at Huf Secure Mobile (Liberkee). You can connect with him on LinkedIn.