mesosphere/marathonArchived
Marathon
Marathon est un orchestrateur de conteneurs Apache Mesos et un gestionnaire de charges de travail distribuées. Il sert de framework pour déployer et gérer des applications conteneurisées et des charges de travail sur un cluster distribué.
Le système fonctionne comme un planificateur de ressources de cluster qui assigne les applications aux ressources disponibles via un plan de contrôle centralisé. Il se distingue par son orchestration basée sur des contraintes de placement d'instances, qui restreint le déploiement des conteneurs en fonction des attributs des nœuds ou des emplacements en rack pour garantir une haute disponibilité.
La plateforme offre des capacités de mise à l'échelle automatique du cluster et de découverte de services conteneurisés pour s'intégrer aux équilibreurs de charge. Elle inclut un système de surveillance de santé utilisant des sondes réseau pour déclencher des processus de récupération automatiques et un système de notification qui diffuse les événements du cluster en temps réel vers des points de terminaison externes.
Les interfaces de gestion sont protégées par authentification et chiffrement SSL.
Features
- Containerized Application Deployment - Orchestrates the deployment and lifecycle of containers across a cluster to maintain scalability and availability.
- Distributed Workload Managers - Acts as a centralized coordinator for deploying, scaling, and monitoring containerized workloads across a distributed Mesos cluster.
- Container Instance Scaling - Adjusts the number of running application instances to match current demand and distribute traffic across the cluster.
- Distributed Application Managers - Coordinates the health and placement of applications across multiple nodes to ensure stability in distributed environments.
- Centralized Control Planes - Implements a centralized control plane to coordinate the lifecycle and scaling of distributed workloads across the cluster.
- Cluster Resource Schedulers - Functions as a controller that assigns applications to available resources based on specific constraints and node locations.
- Docker Container Lifecycle Automation - Automates the lifecycle of Docker and other containers to maintain high availability and scalability.
- Placement Restrictions - Restricts where containers are deployed by matching application requirements against specific node attributes or rack locations.
- Distributed Cluster Management - Provides a centralized control point for coordinating and monitoring the health of applications running across a distributed cluster.
- Distributed Workload Orchestration - Coordinates the placement and health of applications across multiple nodes to ensure stability in distributed environments.
- Mesos Container Orchestration - Deploys and manages Docker containers and other workloads across an Apache Mesos cluster at scale.
- Mesos Container Orchestrators - Acts as a system for deploying and managing containerized applications and workloads across a distributed Apache Mesos cluster.
- Resource Offer Scheduling - Delegates task placement and hardware resource allocation to a distributed systems kernel via a resource offer mechanism.
- Consensus-Based State Replication - Maintains a consistent cluster configuration and application state across multiple nodes using a replicated consensus-based store.
- Traffic Load Balancers - Integrates with proxies and load balancers to distribute incoming network traffic across active application instances.
- Workload Placement Policies - Restricts application instance placement based on physical location to maintain redundancy and prevent service outages.
- Placement Constraints - Restricts where application instances run based on node or rack location to maintain high availability.
- Service Instance Discovery - Tracks active application instances and provides a mechanism for discovering them to enable network traffic distribution.
- Automated Cluster Scaling - Adjusts the number of running application instances based on demand and distributes traffic across available resources.
- Cluster Management APIs - Provides a remote interface for automating application deployments and cluster configuration via scripts and external tools.
- Real-Time Event Notifications - Streams real-time cluster events to external HTTP endpoints to trigger automation and update load balancers.
- Application Health Monitors - Uses HTTP and TCP network probes to verify application availability and trigger automatic recovery processes.
- HTTP Health Probes - Continuously verifies application availability via periodic HTTP network probes to trigger automatic recovery processes.
- Orchestration and Management - Scaling and management for containers on distributed systems.
- System Deployment - Orchestration framework for long-running services on Mesos.
- Workload Orchestration and Scheduling - Container deployment and management on Apache Mesos.
