Implement container monitoring solution for bay

Registered by hongbin on 2016-05-03

Need a container monitoring solution that monitors containers running in the bay and general usage of the bay COE. For example, we could explore adding support for cAdvisor/heapster for this purpose

Blueprint information

Status:
Not started
Approver:
hongbin
Priority:
Medium
Drafter:
hongbin
Direction:
Approved
Assignee:
Cristovao Cordeiro
Definition:
New
Series goal:
Accepted for newton
Implementation:
Unknown
Milestone target:
None

Related branches

Sprints

Whiteboard

Gerrit topic: https://review.openstack.org/#q,topic:bp/container-monitoring,n,z

Addressed by: https://review.openstack.org/407581
    container monitoring in Swarm with Docker v1.10

Gerrit topic: https://review.openstack.org/#q,topic:bp/https,n,z

Gerrit topic: https://review.openstack.org/#q,topic:bug/1646489,n,z

Usage notes:
    In Swarm with Docker v1.10:
     - monitoring is disabled by default. Enable it by setting the label "monitoring_enabled" to True
     - Grafana comes with a default Docker dashboard and default "admin" password. Use "grafana_admin_passwd" label to overwrite it
     - cadvisor and node-exporter will run on all nodes
     - a Prometheus, Fedora and Grafana container stack are then deployed through the Docker API, all in the same node
     - in v1.10 there are no services, killing that node will destroy the stack. When provisioning a new node the stack should be re-created
     - the final Grafana dashboard will be at http://$NODE_running_stack:3000

Addressed by: https://review.openstack.org/426291
    [WIP][k8s] Monitoring with Prometheus and Grafana

(?)

Work Items

Work items:
- container monitoring in Swarm with Docker v1.10 : INPROGRESS
- add cadvisor/prometheus/grafana based monitoring to Kubernetes : INPROGRESS

This blueprint contains Public information 
Everyone can see this information.