fix the hard coded data source problem for rate limit dashboard

#431

remove hard coded project name from dashboard json file

Create new dashboard which calculate the counts & % that rate limits are triggered

#429

Add sql to calculate the counts that rate limits are triggered

Alerts for platform-cluster & monitoring for six-hour mlab-ns clients

Add bq exporter query to monitor mlab-ns 6-hour clients (#422 )
Expands k8s-prometheus scrape job and adds three new k8s-platform alerts (#420 & #421)

Removed prometheus 1.8 config, new alerts and dashboards

Changes include:

• Removed Prometheus 1.8 configuration
• Updated Annotation & Gardener dashboards
• Updated gcp-service-discovery version to v1.3.1
• New alerts:
  • Epoxy server is not online
  • Nodes that fail to boot successfully
  • Too many AppEngine versions
  • Too many inactive AppEngine instances
• Updates to k8s dashboards

Updated Grafana to v5.4.3

This closes m-lab/dev-tracker#94.

Update GMX to 0.1.2

Updated GMX's version number to deploy new version including @nkinkade 's bugfixes.

Rerelease snmp scraping from internal GKE cluster

Merge pull request #394 from m-lab/sandbox-roberto

Change high disk usage threshold for NPAD to 9GB

Revert "Use snmp service running on gke cluster"

Merge pull request #385 from m-lab/sandbox-roberto

Revert "Use snmp service running on gke cluster"

snmp_exporter + kubeIP as k8s deployment

Changes include:

• Added snmp_exporter and kubeIP as Kubernetes deployments
• Added rebot on EB as scraping target
• Several improvements to monitoring and alerts
• Disable collection of gardener traceroute metrics
• Improvements to dashboards

Dashboard improvements, jsonlint and monitoring of SSH on port 22

Changes include:

• Added GMX and lame-duck status for each node in the Ops: Pod overview dashboard
• Added monitoring of SSH running on port 22
• Fixed CPU panel & several improvements to the Pipeline Annotation Service dashboard
• Added jsonlint to .travis.ci