Documenti full-text disponibili:
Abstract
Container orchestration platforms, like Kubernetes, have modernized infrastructure management by automating the deployment, scaling, and recovery of containerized applications. This abstraction has allowed platform and DevOps teams to focus less on manual operations and more on automation and tooling. Internal Developer Platforms (IDPs) have emerged in response to this automation opportunity, encoding DevOps teams' operational knowledge into interfaces developers use directly. At Prometeia S.p.A., an Italian consulting company providing advisory services, research insights and tech solutions, the opportunity was identified to automate the provisioning of ephemeral environments. Ephemeral environments are short-lived deployments generated on-demand and used to validate a small unit of work, usually identified by a pull request. In Prometeia, this type of environment is not automatically provisioned and requires opening a ticket with the DevOps team, which then provisions it manually. To address this opportunity, this thesis presents the design of an IDP whose objective is to automate the provisioning of ephemeral pull request-based environments. Since automated provisioning leads to more environments being created, and therefore to higher infrastructure costs, a second component is introduced to scale environments down when they are not in use. Since covering every possible use case is not feasible, the platform must allow low-level access to the environment definition so developers can customize it to their needs. However, this can introduce misconfigurations that the DevOps team would need to resolve, reintroducing the latency the platform was designed to eliminate. For this reason, a diagnosis component is presented: developers can interact via chat to understand the state of their environment and receive a summarized diagnosis of the problem, along with possible resolutions. The resulting platform is validated against the company's required use cases.
Abstract
Container orchestration platforms, like Kubernetes, have modernized infrastructure management by automating the deployment, scaling, and recovery of containerized applications. This abstraction has allowed platform and DevOps teams to focus less on manual operations and more on automation and tooling. Internal Developer Platforms (IDPs) have emerged in response to this automation opportunity, encoding DevOps teams' operational knowledge into interfaces developers use directly. At Prometeia S.p.A., an Italian consulting company providing advisory services, research insights and tech solutions, the opportunity was identified to automate the provisioning of ephemeral environments. Ephemeral environments are short-lived deployments generated on-demand and used to validate a small unit of work, usually identified by a pull request. In Prometeia, this type of environment is not automatically provisioned and requires opening a ticket with the DevOps team, which then provisions it manually. To address this opportunity, this thesis presents the design of an IDP whose objective is to automate the provisioning of ephemeral pull request-based environments. Since automated provisioning leads to more environments being created, and therefore to higher infrastructure costs, a second component is introduced to scale environments down when they are not in use. Since covering every possible use case is not feasible, the platform must allow low-level access to the environment definition so developers can customize it to their needs. However, this can introduce misconfigurations that the DevOps team would need to resolve, reintroducing the latency the platform was designed to eliminate. For this reason, a diagnosis component is presented: developers can interact via chat to understand the state of their environment and receive a summarized diagnosis of the problem, along with possible resolutions. The resulting platform is validated against the company's required use cases.
Tipologia del documento
Tesi di laurea
(Laurea magistrale)
Autore della tesi
Ceredi, Simone
Relatore della tesi
Correlatore della tesi
Scuola
Corso di studio
Ordinamento Cds
DM270
Parole chiave
Internal Developer Platform,GitOps,Kubernetes,Ephemeral Environments,Kubernetes Operator,Cost Optimization,AI-Assisted Diagnostics
Data di discussione della Tesi
1 Ottobre 2026
URI
Altri metadati
Tipologia del documento
Tesi di laurea
(NON SPECIFICATO)
Autore della tesi
Ceredi, Simone
Relatore della tesi
Correlatore della tesi
Scuola
Corso di studio
Ordinamento Cds
DM270
Parole chiave
Internal Developer Platform,GitOps,Kubernetes,Ephemeral Environments,Kubernetes Operator,Cost Optimization,AI-Assisted Diagnostics
Data di discussione della Tesi
1 Ottobre 2026
URI
Statistica sui download
Gestione del documento: