A flexible and reliable cloud solution for Liantis

Liantis has been supporting starting entrepeneurs and freelancers for almost 100 years with various services. They help their customers with clear information, meticulous administrative support, specific advice and concrete coaching. Liantis’ customers have access to an entrepeneurs’ counter, a social insurance fund, a social secretary and an extensive educational catalog.

The challenge: managing a growing infrastructure

For the last few years, Liantis has been intensively using products from the Atlassian suite. Today, more than 1800 active users fire up these tools every day, cementing the products as an essential part of the daily running of the organization. Because of their ever growing expansion, Liantis was looking for a reliable partner who could manage its infrastructure as well as the applications themselves, allowing Liantis to concentrate on its core business activities without having to worry about their Atlassian stack.

Additionally, because of this continuous growth, the load on applications such as Atlassian Jira kept growing, which could have a negative impact on performance. Liantis therefore wanted to scale up their infrastructure to be able to handle these spikes in usage and keep Jira performance high at all times.

Our solution: Managed Services based on AWS Cloud and Kubernetes

To be able to help Liantis with outsourcing their infrastructure management, we offered our ACA Atlassian Managed Services as a solution. With this solution, we centralize the licenses, hosting, monitoring, infrastructure and maintenance of the applications for a fixed amount per year.

For scaling the applications up or down depending on a rising or declining load, we migrated Liantis’ on-premise Jira appliction to AWS Cloud.

Two Kubernetes clusters for maximal flexibility

As a solution, we proposed to use two different Kubernetes clusters, each with their own AWS account. Splitting the solution up into two clusters allows Liantis to keep the management of rights and costs separate. Additionally, this approach facilitates management and maintenace: whenever there’s no more need for a particular Kubernetes cluster, it can be terminated immediately without impacting other clusters. This also applies to maintenance: if one cluster needs to be taken offline for (scheduled) maintenance, the other can stay online so service is never interrupted.

Reliable infrastructure-as-code

The on-premise infrastructure at Liantis was set up and managed manually. This traditional approach may cause inconsistencies between environments, with a greater risk of human errors.

Through our infrastructure-as-code approach, the entire infrastructure of Liantis is written down in legible code. This way, we’re able to use the same code to quickly set up different environments without running the risk that these environments differ from one another.

Whenever the need arises for new infrastructural components or changes, the code can be quickly and efficiently adjusted to reflect these changes. The new code is then executed in an acceptance environment. When the code is validated, it’s pushed to the production environment. This automatic process drastically reduces the chances of human errors.

Goal of 99.9% availability

The use of containers in a Kubernetes cluster allows us and Liantis to strive towards a rate of availability of 99.9%. Whenever Jira encounters problems, performance hiccups or even flat-out stops working, the current container terminates and a new one automatically spins up. Users don’t even notice this and can just keep working without interruptions. This ‘self-healing’ of Kubernetes clusters also allows an application to recover themselves after office hours. Being unable to continue working or losing data because of crashes are a thing of the past!

Automatic scaling for a more performant application

From the more than 1800 Liantis employees, about 300 use Jira at the same time at any given moment. The combination of infrastructure-as-code and AWS Cloud allows the underlying hardware to simply and efficiently automatically scale up whenever deemed necessary. This way, Liantis doesn’t have to pay for extensive infrastructure when it’s not needed, while still being able to serve all Jira users without delays during peak times.

In the case of any performance problems despite these efforts, our 24/7 Datadog monitoring immediately notifies us of any (potential) problems, allowing us to quickly take action to solve the problem.