specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Etcd
providerId: etcd
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-05'
reconciled: false
tags:
  - FinOps
  - FOCUS
  - Open Source
  - Distributed Systems
  - Kubernetes
description: 'FinOps view of etcd: no vendor invoice. Direct cost is the underlying compute, storage
  (low-latency SSD), and networking required to run a quorum-typically-3 cluster. When etcd is consumed
  via managed Kubernetes (EKS, GKE, AKS), its cost is bundled into the control-plane fee on the cloud
  bill.'
sources:
  - https://etcd.io/
  - https://www.cncf.io/projects/etcd/
notes: No vendor invoice from the etcd project. FinOps surface is infrastructure cost (self-hosted) or
  a portion of managed-Kubernetes control-plane spend.
alignedWith:
  framework: FinOps Foundation Framework
  frameworkUrl: https://www.finops.org/framework/
  dataSpec: FOCUS
  dataSpecVersion: '1.3'
  dataSpecUrl: https://focus.finops.org/focus-specification/v1-3/
publisherName: Cloud Native Computing Foundation
serviceCategory: Distributed Systems / Key-Value Store
billingModel:
  pricingCategory: Open Source (no direct cost)
  billingFrequency: N/A
  billingCurrency: USD
  chargeCategories:
    - Usage
focusColumns:
  ServiceName: etcd
  ServiceCategory: Distributed Systems / Key-Value Store
  ProviderName: etcd
  PublisherName: Cloud Native Computing Foundation
  InvoiceIssuerName: N/A (open source)
  BillingCurrency: USD
meters:
  - name: compute_hours
    unit: instance-hour
    aggregation: sum
    dimensions:
      - cluster
      - region
      - instance_type
  - name: ssd_storage
    unit: GB-month
    aggregation: max
    dimensions:
      - cluster
      - region
  - name: backup_storage
    unit: GB-month
    aggregation: max
    dimensions:
      - cluster
  - name: network_egress
    unit: GB
    aggregation: sum
    dimensions:
      - cluster
      - region
principles:
  - name: Visibility
    description: Scrape etcd's Prometheus metrics (mvcc_db_total_size_in_bytes, disk fsync latency, leader
      changes) and join with cloud cost data per cluster.
  - name: Allocation
    description: Tag etcd nodes/disks per cluster and per environment so platform-team cost is attributable
      to consuming Kubernetes/control-plane workloads.
  - name: Optimization
    description: Co-locate etcd peers on low-latency NVMe SSD; right-size CPU/memory; use defrag and
      compaction to keep DB small; consolidate small clusters where multi-tenant model permits.
  - name: Accountability
    description: Platform team owns etcd cluster cost; charge back to consuming Kubernetes control-plane
      tenants on a per-cluster or per-namespace basis.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com