- Aggregate by container where possible.
- Correct the disk I/O queries.

We're planning on removing the priority and weight keys and ordering
panels and groups in the order that they are defined.

- Add a new out-of-the-box dashboard for monitoring K8s pod health.

- Sort dashboards in the dashboard selector dropdown by their
display_name.

References are changed in:
  - FE code & tests
  - BE code & tests
  - docs & docs screenshots
  - translation files

- K8s 1.14 introduced the pod and container labels and deprecated the
pod_name and container_name labels. K8s 1.16 removes the pod_name and
container_name labels.

- Metrics from K8s versions before 1.14 contain pod_name and
container_name labels. Metrics from K8s 1.14/1.15 contain pod, pod_name,
container and container_name labels. Metrics from K8s 1.16 onwards
contain pod and container labels.

- According to our docs, we need to support K8s 1.12 to 1.16.

- This commit changes existing queries to use pod instead of pod_name
and container instead of container_name. These changed queries should
work on K8s 1.14 onwards.

- This commit also adds a second query using `OR`. The second query
uses pod_name instead of pod, and container_name instead of container.
This second query should work on K8s 1.12 to 1.15.

AS next step towards moving cluster health dashbaord to gitlab
core, we need to move cluster metrics services to core

Add an instance variable that allows users to filter the resource
usage charts by the instance label.

Add charts showing the CPU and memory usage of each instance to the
self monitoring default dashboard.

- Use {{ }} syntax for variable substitution in common_metrics.yml
since we've deprecated the %{ } syntax.

Add `format` key to the panels is the default dashboard to format data.

Add documentation to match the new changes for users.

Add a pod metrics dashboard that can be used to display pod health
charts. The queries use a variable called pod_name which will be
replaced with the actual pod_name before the query is sent to
Prometheus.

Adds a time series component for line and area charts.
Displays new charts in the dashboard.

- Use dynamic components for line/area swapping
- Add new line charts to dashboard in 2 panels

This MR updates the y_label values and the label
values in common_metrics and cluster_metrics to
append the units of measurement to the labels.

These aren't used in CE, but it's safe to include them.

Dashboards will be ordered in descending order of priority. The word
'priority' seems to signify descending order while 'order' signifies
ascending order.

Dashboards will be ordered in descending order of priority. The word
'priority' seems to signify descending order while 'order' signifies
ascending order.

- They are not used by the frontend.

- Change it to the new dashboard syntax described here:
https://gitlab.com/gitlab-org/gitlab-ce/issues/60383.

- They are not used by the frontend.

- Change it to the new dashboard syntax described here:
https://gitlab.com/gitlab-org/gitlab-ce/issues/60383.

Reorders the chart queries
Adds line thickness

Render both usage and capacity in chart