# koordinator-sh/koordinator

A QoS-based scheduling system brings optimal layout and status to workloads such as microservices, web services, big data jobs, AI jobs, etc.

Repository: https://github.com/koordinator-sh/koordinator
Canonical: https://ross.abutalabs.com/products/koordinator
Homepage: https://koordinator.sh
Language: Go
License: Apache-2.0
License Family: permissive
Topics: kubernetes, colocation, scheduling, cloud-native, cost-optimization, workload-consolidation
Last push: 2026-08-26T14:17:44+00:00

## Health v2 (maintenance only)
Score: 84/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 55, longevity 100
- inputs: {"age_days": 1631, "days_push": 7, "days_rel": 140, "gap_med": 166, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1745, forks 477 (observed 2026-08-28T04:05:30.598745+00:00)

## What it is
Koordinator is a QoS-based scheduling system for Kubernetes that colocates latency-sensitive microservices with batch, big data, and AI workloads. It improves resource utilization through overcommitment, fine-grained resource isolation, load-aware scheduling, and descheduling while preserving QoS guarantees.

## Use cases
- improve kubernetes cluster resource utilization
- colocate batch jobs with latency-sensitive microservices
- run spark and tensorflow jobs alongside web services on kubernetes
- reduce cloud costs by consolidating workloads
- schedule pods based on node load and QoS
- overcommit resources while protecting production SLOs
- evict and rebalance pods when nodes are overloaded

## When to choose
- you run mixed latency-sensitive and batch workloads on Kubernetes and want higher utilization
- you need QoS-aware, load-aware scheduling and resource isolation beyond native Kubernetes
- you want proven production-grade colocation with priority and overcommitment mechanisms

## When to avoid
- you run a small cluster with low utilization where native Kubernetes scheduling suffices
- you cannot run Linux kernel 4.19+ or deploy DaemonSet node agents
- you need a scheduler for non-Kubernetes environments

## Facets
- artifact type: service
- maturity: active
- function: container-orchestration, scheduling, monitoring
- domain: cloud-computing, microservices
- platform: go, self-hosted
- tags: qos, colocation, resource-utilization, workload-consolidation, cost-optimization, kubernetes-scheduler, descheduler, containers, devops, kubernetes, linux, docker

## Member repositories
- koordinator-sh/koordinator (main) score 84

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:30.598745+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:29:14.679251+00:00, confidence not recorded.
  - readme: https://github.com/koordinator-sh/koordinator (fetched 2026-08-28T04:05:30.598745+00:00, sha 205ae500e7c0)
  - homepage: https://koordinator.sh (fetched 2026-08-29T11:07:10.900825+00:00, sha cbb96af8533f)
  - site_page: https://koordinator.sh/docs (fetched 2026-08-29T11:07:10.910081+00:00, sha 4d7f268eb66c)
  - site_page: https://koordinator.sh/docs/next (fetched 2026-08-29T11:07:10.912181+00:00, sha f33f0a655c8f)
  - site_page: https://koordinator.sh/docs/v1.7 (fetched 2026-08-29T11:07:10.914059+00:00, sha 94edc32c0169)
  - site_page: https://koordinator.sh/docs/v1.6 (fetched 2026-08-29T11:07:10.915805+00:00, sha b356f667d926)
  - site_page: https://koordinator.sh/docs/installation (fetched 2026-08-29T11:07:10.917563+00:00, sha 26eaaadab020)
  - site_page: https://koordinator.sh/docs/user-manuals/colocation-profile (fetched 2026-08-29T11:07:10.919712+00:00, sha 56fc01afab07)
  - site_page: https://koordinator.sh/docs/architecture/overview (fetched 2026-08-29T11:07:10.921397+00:00, sha e93a63202c90)
- Data as of 2026-08-30T08:39:29.467469+00:00.