DOCSEARCHbyAlgolia
Submit Docs

Atmosphere by Vexxhost

repository·main·Indexed 41 days ago

https://github.com/vexxhost/atmosphere

Infrastructure and cloud-native tooling managed by Vexxhost, featuring Helm charts for ceph-csi-rbd, CoreDNS, and FRR-K8s. Includes documentation for deploying RBD volume support, configuring CoreDNS deployment modes and autoscaling, and managing FRR-K8s Custom Resource Definitions (CRDs).

Tokens
295.7K
Snippets
647
Records
938
Agent score
51%

What's inside Atmosphere

  1. Overview of the Loki Helm Chart

    main

    The loki Helm chart provides deployment capabilities for Grafana Loki and Grafana Enterprise Logs. It supports three deployment modes:

    1. Simple mode: For smaller workloads.
    2. Scalable mode: For increased throughput.
    3. Distributed mode: For highly scalable, microservices-based architectures.

    This chart is a wrapper/deployment for the official Loki project.

  2. Overview of the Emulex LightPulse Fibre Channel (LPFC) driver role

    main
    The lpfc role is designed to tune the Emulex LightPulse Fibre Channel (LPFC) driver specifically for OpenStack environments. Use this role to ensure the driver is optimized for the performance and stability requirements of OpenStack workloads.

  3. Overview of Node Feature Discovery (NFD)

    main
    Node Feature Discovery (NFD) is a Kubernetes add-on designed to detect hardware features and system configurations on cluster nodes. Once detected, these features are advertised as Kubernetes node labels, allowing workloads to be scheduled based on specific hardware capabilities (e.g., CPU instructions, GPU presence, or specific kernel configurations). NFD is extensible and supports various vendor-specific and application-specific labeling requirements.

  8. Access Grafana and Prometheus UIs

    main

    Grafana Access

    Grafana is exposed via an Ingress using the kube_prometheus_stack_grafana_host variable. Authentication is handled by Keycloak using the following roles:

    • grafana:admin: Full access to all resources (dashboards, users, teams).
    • grafana:editor: Can view and edit dashboards, folders, and playlists.
    • grafana:viewer: Can view dashboards, playlists, and query data sources.

    Prometheus Access

    Prometheus is exposed via an Ingress using the kube_prometheus_stack_prometheus_host variable. By default, it is protected by an oauth2-proxy service for authentication.

  12. Use the `reboot` operational role

    main

    The reboot role is an operational tool designed to perform a safe, automated reboot of a server. It follows a specific lifecycle to ensure monitoring systems do not trigger false alarms during the downtime:

    1. Safety Checks: Runs a predefined set of checks to verify the server is in a safe state to reboot.
    2. Silence Creation: If checks pass, it automatically creates a silence in the monitoring system to prevent alarms from firing.
    3. Reboot Execution: Initiates the server reboot.
    4. Silence Removal: Once the server is back online and reachable, the role removes the silence to resume normal monitoring.