What's Changed

• Bug Fixes
• Update templates
• Unified configuration files to one

What's Changed

• Custom exporter via Moneo CLI
• GPU sample rate is now adjustable
• Update to Moneo docker. Moneo exporters launched within docker
• Refresh Linux server deployment and docs
• Addition frontend network naming arguments for node exporter

Full Changelog: https://github.com/Azure/Moneo/compare/v0.3.3...v0.3.4

What's Changed

• Doc updates
• Arm templates for Managed Prometheus and Grafana infrastructue
• slurm example updates
• bug fixes

Full Changelog: https://github.com/Azure/Moneo/compare/v0.3.2...v0.3.3

minor bug fixes

What's Changed

Add physical hardware hostname to managed Prometheus deployment method.

What's Changed

• Managed Prometheus/Grafana deployment method set to preferred method for Services
• IB Link flap metric fixes
• Bug fixes

What's Changed

• Managed Prometheus Agent Integration
• fix bug related to accelerated network
• Alma bugfix

Moneo no longer uses Ansible and is replaced with pssh
Moneo can now be launched as a Linux service
Moneo can now publish to Geneva or Azure monitor/Log analytics.
Small bug fixes

What's Changed

• Geneva cert authentication feature
• Doc updates
• Bug fixes

What's Changed

• Moneo no longer uses Ansible and is replaced with pssh
• Moneo can now be launched as a Linux service
• Moneo can now publish to Geneva or Azure monitor/Log analytics.
• Small bug fixes

What's Changed

• Moneo no longer uses Ansible and is replaced with pssh
• Moneo can now be launched as a Linux service
• Moneo can now publish to Geneva or Azure monitor/Log analytics.

What's Changed

• Logging to CLI and Exporter output to files:
  • CLI log location: Moneo/ moneoCLI.log
  • Exporter log location: on worker node: /tmp/moneo-worker/moneoExporter.log
• Adding XID and Link Flap graphs to capture when these issues occur

What's Changed

• Hotfix for Node Exporter Job ID Update by @yakovdyadkin in https://github.com/Azure/Moneo/pull/33
• [CI/CD] Add Moneo CI Pipeline by @RyoYang in https://github.com/Azure/Moneo/pull/28
• Adding initial components for resource filtering by @rafsalas19 in https://github.com/Azure/Moneo/pull/24
  • A future roadmap component

What's Changed

• Features
  • Node exporter for metrics with lower sample rate
  • Ansible deployment can be parallelized via Moneo CLI
• Addition of new metrics
  • CPU, Memory, and front-end network telemetry added
  • Additional metrics added to cluster view dashboard
• Fixes and other changes
  • Pass in IB counter file as argument for IB net exporter.
  • Remove Python 2 dependencies
  • Fixes to the job update for Nvidia exporter
• Updating documentation
  • Node/Base exporter examples of custom telemetry addition
  • Examples of Slurm prolog/epilog integration

Moneo is a distributed GPU system monitor for AI workflows.

Moneo orchestrates metric collection (DCGMI + Prometheus DB) and visualization (Grafana) across multi-GPU/node systems. This provides useful insights into workflow and system level characterization.

What's Changed

• Features
  • Moneo CLI added to deploy, shutdown, and update Moneo workflow
  • Cluster wide dashboard added to provide overall view of cluster
  • Job filtering at node level
  • Update AMD exporter (Experimental)
  • Azure Application Insights Integration (Experimental)
• Addition of new metrics
  • IB port metrics to observe link and port status
  • GPU throttling metrics
• Fixes and other changes
  • Profiling Metrics disabled by default to avoid additional overhead
  • Skip DCGM install if already installed
• Updating documentation
  • Quick start guide
  • Job filtering document

• Fix for net exporter due to python version issues
• added additional data (requirements and access) to the readme

• Fixed hostname resolution issues allowing this now to work with Azure Cycle Cloud.
• Fixes to account for shared home directory.
• Fixed a small net exporter issue

Moneo v0.1.0 Release Notes

Moneo GPU Monitoring Framework

• First release of Mone, includes ansible framework to coordinate docker containers and metric collection workers.
• Grafana front end implementation.
• Includes Prometheus DB to store metrics.
• Reference ReadMe doc for usage instructions.