FinOps y Sostenibilidad Cloud: Un Equilibrio Entre Eficiencia Financiera y Responsabilidad Ambiental - Jose María Flores Zazo(Solutions Architect @ Avanade)
Una pequeña charla de 5 minutos que muestra los tips que hacen diferente a FinOps de Sosntenibilidad. Actualmente veo como se usa FinsOps para vender sostenibilidad cuando no debería ser así.
Cost-based strategy to sustainability at Grafana - Paula Julve(Senior Software Engineer @ Grafana Labs)
There's been an increased focus on sustainability, but oftentimes money speaks louder. So, let's bring them together.
At Grafana, we're using kube-state-metrics and OpenCost to monitor how much money we're spending on resources sitting idle in our k8s clusters. With this data in hand, we've been able to leverage Karpenter to make the most sustainable choices that will still make sense for the business.
Cost Optimizations for PostgreSQL clusters running on Kubernetes -Dimitris Kapanidis(Founder @ Harbur Cloud Solutions)
This is a lightning talk about cost optimizations (and subsequently more greener solutions) of PostgreSQL instances running on top of Kubernetes instances by measuring the actual CPU, memory and disk usage, and optimizing the resource requests and limits to the clusters, along with other tweeks such as enabling savings plans on AWS, switching to more newer and greener CPU types, autoscaling clusters, treating prod and non-prod differently and more.
Being an energy-efficient data center with EAR - Jalal Lakhlili(HPC Engineer @ Energy Aware Solutions S.L.)
EAR is an open source system software for energy management, monitoring and optimization in HPC data centers. It has been developed by BSC and since 2020 it has been commercialized by Energy Aware Solutions, a BSC- UPC spinoff. EAR includes in a single software most of the features required by data centers in terms of hardware monitoring, job accounting, energy management and cluster powercap. All the EAR components collaborate together to guarantee the job power consumption is according to job characteristics (and no more), and the total power consumption is below operational limits. During the talk we will present main EAR features in the previous topics and our current and short term plans to extend EAR to include other parts of the DC, other types of use cases, other architectures, etc. All of these extensions addresses the common goal of making energy efficient data centers a reality both from hardware and software perspectives.
In the evolution of provisioning resources for our software applications, the shift towards cloud-native adoption has brought significant changes. As developers, it is paramount for us to cultivate an efficiency-oriented mindset when utilizing cloud services. In light of the prevailing global circumstances, there arises a pressing need for a set of principles and tools that can effectively bridge the gap between utilizing cloud resources and embracing more sustainable practices.
Carbon aware time shifting with KEDA - Ross Fairbanks(Platform Engineer @ FlatPeak Energy)
In this session, we will look at what carbon intensity is, why it varies depending on how much renewable energy is available, and how we can shift demand by time and location to reduce carbon emissions. This will be followed by a demo of using the carbon-aware-keda-operator from Microsoft to do demand shaping using carbon intensity data from WattTime.