feat: Kubernetes Logging Monitoring and Alerting Building Blocks

[andrecolin]

There will be many solutions to monitoring Kubernetes and the Applications running on Kubernetes. This issue request is to identify the components that can easily be combined into a Deployment Helm Chart and use these building blocks to form several best of breads Logging Monitoring and Alerting (LMA) Solution.
We need the following building blocks Helm Charts making sure they are idempotent and can be combined with any of the other Charts to form a complete LMA Solution. Some on the list may already have Charts, we just need to make sure it is in the correct format we desire. Part I will be to Helm Chartify the components and part 2 will be to configure sample deployment charts to make a solution

ElasticSearch
Kibana
Heapster
Grafana
InfluxDB
FluentD
Prometheus
OpenTDB
SNAP

[v1k0d3n]

@andrecolin it losses some value without the links! :)

you mentioned having charts for each of these items. do these work provide good starters? i would imagine that many of these items we want to really be upstream in other sources (when possible):

ElasticSearch: kubernetes/charts/incubator/elasticsearch, or there are 3 versions of elasticsearch in fabric8io's repo. there is also Samsung's k2 Charts that contain's an elasticsearch chart.
Kibana: Same links above have Kibana versions.
Heapster: Samsung k2 has heapster
Grafana: Grafana is in all of them.
InfluxDB: Kubernetes
FluentD: Fabric8 has a FluentD chart to use
Prometheus: Fabric8 has 2 versions, Samsung k2 has one, and Kubernetes Charts has one.
OpenTDB: New Item
SNAP: New Item

[andrecolin]

We will evaluate each Chart & Combined Deployment to see if it already a good fit. If we see benefits to tweak a chart to make it more universal, then we can ask chart authors to make the tweaks else we can ask the community to help with the chart

[v1k0d3n]

@andrecolin we may be able to see if the authors would accept a PR for additional functionality. if they are specific cases (like faboric8; they have a PaaS offering built around their chart repo), they may not want too...in which we can add to our internal repo. this may end up being an att-comdev/charts repo...we just don't know quite yet.