Proposed Feature 01 Monitoring and Managing Cloud Usages in ACS C O M M I T E D D I G I T A L E X P E R T I S E D I M S I . F R
Your Speake rs Grégoire LAMODIERE CEO, Cloud Solutions Architect Joffrey LUANGSAYSANA Cloud Expert 02
03 Who are we? About Dimsi. Why a reporting feature? Current trends, clients feedback State of the art Reporting in Azure What about Cloudstack Usage Service / instances consumption history Proposed Feature Explanation / mockup Let's chat An open discussion
04 25 experts on three locations Team Size Lorient Paris Lyon Creation 2006 Practices
05 EXPERTISE Microsoft Power Platform Cloud & Services Modern Apps
06 Why a reporting feature #1 ? Many organizations are building roadmaps re-thinking their infrastructure practices - Onpremise : VMWare clients are no longer confident regarding Broadcom sales strategy => considering technology migration o Virtual Machines only : Proxmox, XCP, KVM => CS as KVM management o Complex workloads / Orchestrations : Cloudstack, Openstack, OpenNebula and others - Cloud : - Costs are much higher than expected + digital sovereignty => Which can lead to Cloud hybrid or full repatriation - Global : Many opportunities for Apache Cloudstack adoption
07 Why a reporting feature#1 ? From our leads feedbacks: 3 barriers to Apache CloudStack Adoption Barriers Customers Solution Migration path « This migration project will need massive effort from my team » ✅ Simplified by mutiple tools (ie. CloudStack VM Ingestion, virt-v2v) Knowledge « My team has only basic knowledge in Linux world, I’m not sure they’ll be able to handle this new infrastructure » ✅ Bootcamp training services ✅ Design and implementation services ✅ Support services Reporting « I need an infrastructure report, and my management is pushing to go for greenops. They love our current cloud reporting system » ❌ No standard reporting capabilities in cloudstack ❌ No built-in feature to define consumptions KPIs (electrical, turnover, etc) ❌ Even usage service is only API Based, with no Frontend
08 State of the art: Example - Microsoft Azure dashboard
09 State of the art: Microsoft GreenOps Dashboard Goal : Get the CO2 emissions of services provided Pros : very simple agragated view Cons : Not really accurate, mainly based on service location
10 What about Cloudstack Data • Usage service with API capabilities: Instances and infrastructure consumption • listCapacity API command: Capacity consumption by pod or cluster • listVirtualMachinesUsageHistory • listSystemVmsUsageHistory • listUsageRecords listUsageRecords startdate=2025-03-01 enddate=2025-03-02 domainid=9d190e08-f1b0-412d-b28d-811e61191a9e { "account": "Admin", "accountid": "02383f45-f318-439a-b8d9-6addbe7aa", "cpunumber": 4, "cpuspeed": 2400, "description": "Running VM usage for SRV1 (i-12-18680-VM) (bf83ef25-e2e2-432b-8e53-49e5ef4d718f) using service offering KVM (4104c50c- fb7a-45cb-a788-dec8d39a2d7a) and template Win-Srv-2016-Std-KVM (9405a5e3-5460-4e9c-a773-b95dc17a4346)", "domain": « 387", "domainid": "9d190e08-f1b0-412d-b28d-811e61191a9e", "domainpath": "/IAAS/387/", "enddate": "2024-03-01'T'23:59:59+00:00", "memory": 4096, "name": « SRV1", "offeringid": "4104c50c-fb7a-45cb-a788-dec8d39a2d7a", "oscategoryid": "47c7f725-1fed-11e8-a604-f61f8a45771f", "oscategoryname": "Windows", "osdisplayname": "Windows PV", "ostypeid": "47f03dcf-1fed-11e8-a604-f61f8a45771f", "rawusage": "24", "startdate": "2024-03-01'T'00:00:00+00:00", "tags": [], "templateid": "9405a5e3-5460-4e9c-a773-b95dc17a4346", "type": "KVM",
11 What about Cloudstack Reporting & UI • CS 4.20 : usage-ui to display and download usage records • Cluster / host / VM ressources metrics
12 Proposed feature Expand Cloudstack reports and dashboarding for both providers and end-users • Ressources & Ops : • Dashboard solution for consumption services with history • Provider and customer view • Filtering : infrastructure, domain, account, projects • GreenOps : • Generate greenops dashboard • Set solution to gather electrical consumption from hosts (automatic / manual) • Identify impact of instance in host • Identify impact of a network in the networking insfrastructure • Set consumption based on localisation
13 Proposed feature Ressources & Ops Dashboard: Global Admin or IT Service  Number of ressources created and deleted (Domains, VPC, Networks, VMs)  Network and Storage  Number of logins, Powered off / on VMs
14 Proposed feature Ressources & Ops Dashboard: End user Least used VMs in % consumption for set period
15 Proposed feature
16 Limitations We anticipate Performances : • Limitations for larger infrastructure • Time needed to generate report / dashboard • Data history processing Technical : o Ops • Retrieve %use for some OS • DB performances impact and storage for metrics history for VMs and hosts o GreenOps • Find generic solution to gather host real consumption • Core infrastructure (switching, etc) : real consumption ? • Component impact, manufacturing impact and production resources
Thanks For Your Attention Let’s chat

Proposed Feature: Monitoring and Managing Cloud Usage Costs in Apache CloudStack

  • 1.
    Proposed Feature 01 Monitoring andManaging Cloud Usages in ACS C O M M I T E D D I G I T A L E X P E R T I S E D I M S I . F R
  • 2.
    Your Speake rs Grégoire LAMODIERE CEO, Cloud SolutionsArchitect Joffrey LUANGSAYSANA Cloud Expert 02
  • 3.
    03 Who are we? AboutDimsi. Why a reporting feature? Current trends, clients feedback State of the art Reporting in Azure What about Cloudstack Usage Service / instances consumption history Proposed Feature Explanation / mockup Let's chat An open discussion
  • 4.
    04 25 experts onthree locations Team Size Lorient Paris Lyon Creation 2006 Practices
  • 5.
  • 6.
    06 Why a reportingfeature #1 ? Many organizations are building roadmaps re-thinking their infrastructure practices - Onpremise : VMWare clients are no longer confident regarding Broadcom sales strategy => considering technology migration o Virtual Machines only : Proxmox, XCP, KVM => CS as KVM management o Complex workloads / Orchestrations : Cloudstack, Openstack, OpenNebula and others - Cloud : - Costs are much higher than expected + digital sovereignty => Which can lead to Cloud hybrid or full repatriation - Global : Many opportunities for Apache Cloudstack adoption
  • 7.
    07 Why a reportingfeature#1 ? From our leads feedbacks: 3 barriers to Apache CloudStack Adoption Barriers Customers Solution Migration path « This migration project will need massive effort from my team » ✅ Simplified by mutiple tools (ie. CloudStack VM Ingestion, virt-v2v) Knowledge « My team has only basic knowledge in Linux world, I’m not sure they’ll be able to handle this new infrastructure » ✅ Bootcamp training services ✅ Design and implementation services ✅ Support services Reporting « I need an infrastructure report, and my management is pushing to go for greenops. They love our current cloud reporting system » ❌ No standard reporting capabilities in cloudstack ❌ No built-in feature to define consumptions KPIs (electrical, turnover, etc) ❌ Even usage service is only API Based, with no Frontend
  • 8.
    08 State of theart: Example - Microsoft Azure dashboard
  • 9.
    09 State of theart: Microsoft GreenOps Dashboard Goal : Get the CO2 emissions of services provided Pros : very simple agragated view Cons : Not really accurate, mainly based on service location
  • 10.
    10 What about Cloudstack Data •Usage service with API capabilities: Instances and infrastructure consumption • listCapacity API command: Capacity consumption by pod or cluster • listVirtualMachinesUsageHistory • listSystemVmsUsageHistory • listUsageRecords listUsageRecords startdate=2025-03-01 enddate=2025-03-02 domainid=9d190e08-f1b0-412d-b28d-811e61191a9e { "account": "Admin", "accountid": "02383f45-f318-439a-b8d9-6addbe7aa", "cpunumber": 4, "cpuspeed": 2400, "description": "Running VM usage for SRV1 (i-12-18680-VM) (bf83ef25-e2e2-432b-8e53-49e5ef4d718f) using service offering KVM (4104c50c- fb7a-45cb-a788-dec8d39a2d7a) and template Win-Srv-2016-Std-KVM (9405a5e3-5460-4e9c-a773-b95dc17a4346)", "domain": « 387", "domainid": "9d190e08-f1b0-412d-b28d-811e61191a9e", "domainpath": "/IAAS/387/", "enddate": "2024-03-01'T'23:59:59+00:00", "memory": 4096, "name": « SRV1", "offeringid": "4104c50c-fb7a-45cb-a788-dec8d39a2d7a", "oscategoryid": "47c7f725-1fed-11e8-a604-f61f8a45771f", "oscategoryname": "Windows", "osdisplayname": "Windows PV", "ostypeid": "47f03dcf-1fed-11e8-a604-f61f8a45771f", "rawusage": "24", "startdate": "2024-03-01'T'00:00:00+00:00", "tags": [], "templateid": "9405a5e3-5460-4e9c-a773-b95dc17a4346", "type": "KVM",
  • 11.
    11 What about Cloudstack Reporting& UI • CS 4.20 : usage-ui to display and download usage records • Cluster / host / VM ressources metrics
  • 12.
    12 Proposed feature Expand Cloudstackreports and dashboarding for both providers and end-users • Ressources & Ops : • Dashboard solution for consumption services with history • Provider and customer view • Filtering : infrastructure, domain, account, projects • GreenOps : • Generate greenops dashboard • Set solution to gather electrical consumption from hosts (automatic / manual) • Identify impact of instance in host • Identify impact of a network in the networking insfrastructure • Set consumption based on localisation
  • 13.
    13 Proposed feature Ressources& Ops Dashboard: Global Admin or IT Service  Number of ressources created and deleted (Domains, VPC, Networks, VMs)  Network and Storage  Number of logins, Powered off / on VMs
  • 14.
    14 Proposed feature Ressources& Ops Dashboard: End user Least used VMs in % consumption for set period
  • 15.
  • 16.
    16 Limitations We anticipate Performances : •Limitations for larger infrastructure • Time needed to generate report / dashboard • Data history processing Technical : o Ops • Retrieve %use for some OS • DB performances impact and storage for metrics history for VMs and hosts o GreenOps • Find generic solution to gather host real consumption • Core infrastructure (switching, etc) : real consumption ? • Component impact, manufacturing impact and production resources
  • 17.

Editor's Notes

  • #3 Joff
  • #4 Joff
  • #5 Joff
  • #6 Greg he cloud's promises have not been kept. Learn the hard way Mouvement de fond des entreprises pour repenser leurs pratiques Cloud Pour celles qui sont restées onprem : VMWare Exode, besoins qui se complexifient Pour celles qui sont dans le cloud :
  • #7 Greg : Dans nos séries d’AVV, nous avons identifié plusieurs barrières à contourner pour faciliter l’adaption CS En conclusion de cette slide : aujourd’hui, nous allons nous focaliser sur le sujet pour lequel nous ‘navons pas de réponse à apporter : le reporting
  • #8 Greg
  • #9 Greg Même si on sait que c’est beaucoup de bullshit, c’est un reporting grandement apprécié des client
  • #10 Joff
  • #11 Joff
  • #12 Joff puis Greg
  • #13 Joff
  • #14 Joff
  • #15 Greg
  • #16 Joff & Greg retention: the value of the configuration vm.stats.max.retention.time; interval: the value of the configuration vm.stats.interval; https://docs.cloudstack.apache.org/en/latest/adminguide/virtual_machines.html