As part of its data centre exit strategy, Veolia has migrated all of its data centres to the AWS Cloud. This technological transformation has enabled an organisational transformation in which business teams have become autonomous in their developments and can focus on providing value-added functions. To extend this approach even further, the platform domain has taken over the internal operation of its AWS cloud infrastructure and adopted an observability approach by automating its service offerings.
Sébastien Dorneval, Director of Platforms at Veolia Water France, talks about this project with Devoteam.
The challenge
Can you present the context of the project?
When we started the project, we had migrated all of our data centres to the AWS Cloud and more than 90% of the migration had already been completed. After this migration, with the advent of the use of managed infrastructure services, the question arose as to the interest in maintaining the outsourcing of our infrastructure. Especially since our teams had reached a good level of maturity in terms of tooling and automation (Terraform, Ansible). These were topics on which we worked with Devoteam.
In parallel, we also wanted to introduce an observability approach and go one step further in monitoring the system. In this context, we have decided to move from outsourcing to the operation of our infrastructure, again with the aim of optimising the performance of the cloud, modernising our working methods and offering tailor-made solutions to business teams.
The solution
How did Devoteam support you in this project?
Our business domain teams are autonomous in their construction and operation, and we provide them all the common services. For some teams, we also offer the services for which they do not have the time or competence to develop, so these services must be as automated as possible.
Devoteam led the project and contributed to the automation of the service offerings and the design of the tools. For example, patch management automation is one of the services we offer to reduce some of the Ops load automatically. These automated services allowed us to take over the managed perimeter completely, but with a small team. We also worked with Devoteam to build a fully automated alerting and monitoring chain around the Dynatrace tool, on which we had already performed an observability POC. What interested us, was that in addition to APM (Application Performance Management), Dynatrace also enables monitoring and provides an AI-based learning system for alarm recommendations.
The second point that convinced us, was the strong integration into our existing tools. We were able to integrate metrics from SaaS offerings as well as those of our ELB or EC2 on AWS. Dynatrace also has the advantage of offering a SaaS interface that we can share with all teams. This centralised vision also allows us to avoid silos due to specific business processes.
What is the benefit of this project?
Previously, supervision was only visible through outsourcing reporting. We have now regained control of our data. In addition, the implementation of Dynatrace and OpsGenie allowed us to automate Level 1 alerting and automatically manage calls and escalations with ticket assignment in ServiceNow.
Is this a step towards a more advanced observability phase?
Yes, regaining control of our infrastructure was a first step in creating a monitored and automated vision of applications. We will now be able to work on observability and take advantage of Dynatrace’s APM capabilities. Ultimately, the goal is to deliver a consolidated vision of our IT, centralise operational data on SLAs, monitor it, and share it with as many people as possible. We will also highlight other indicators from the cloud, with the idea being to observe both the infrastructure and the applications.
We have gained autonomy, agility and the ability to work in a team while reducing infrastructure costs.
The added value
What was the added value of the Devoteam intervention?
Devoteam gave us a native vision of automation from the beginning of the project. More specifically, the team contributed to three points: automating patch management, monitoring and automating the incident solution, and managing the entire project. This included working with suppliers and a timing that brought many challenges:
- Predicting the release of the tool
- Performing a double run with the previous tool
- Management of validation constraints at group level
- Accompanying the run of the new tool while our teams have acquired the competence
Did migrating all data centres to the AWS Cloud meet your expectations?
Yes, we have gained autonomy, agility and the ability to work in a team while reducing infrastructure costs. The move to the cloud has re-evaluated the work of the teams that are now being mobilised for higher value-added topics because they no longer waste time running the infrastructure. The cloud has also brought us autonomy and control in our deployments, we are no longer afraid to go into production. In general, the cloud is a transformation. We had planned an organisational transformation in agile mode, with teams managing both the build and the run, but this transformation would not have been possible if it had not been accompanied by a technological transformation. To support teams in their developments, we needed a cloud provider that would enable this flexibility and delegation, and that’s exactly what the AWS Cloud brought us.
