Fixes missing metric alert

This is provisional release to address a production alert which has already been identified and fixed in staging. It addresses this PR: #611

It also brings along with it removal of the downloader cluster deployments from the Travis config.

Weekly release

• Deploy a new gcs-exporter to the prometheus-federation cluster.
• Renames the data-processing cluster, along with any config changes to facilitate the rename.
• Adds a new alert for missing "generic" (non-experiment specific) metrics upon which mlab-ns relies.
• Properly prefixes mlab-ns-related alerts with MlabNS_.

Weekly release

Further reduce number of alerts with severity=page (#602)

* Further reduce number of alerts with severity=page.

* PlatformCluster_FederationScrapeJobFailing has severity=ticket now.

Weekly release

Rollback bqx version (#601)

* Rollback bqx version

* Fix project flag

Weekly release

Point of interest in this release:

• Removes virtually all legacy/PLC configurations.
• Updated versions for most all Docker images for prometheus-federation cluster.
• New alert for when the Prometheus VM's persistent disk gets too full.
• A new tcpinfo dashboard.
• A couple small dashboard fixes/improvements.

Update neubot targets and Grafana dashboards

Add pusher dashboard (#572)

* Add new alert dashboard

* Add reference to pusher dashboard

Use cert-manager, update GMX and alerts

v2.28.0

Remove unused bqx queries (#568)

Weekly release

• An updated version of GMX: v1.1.0
• Two small fixes/improvements to a couple of Grafana dashboards.

Weekly release

• Updates the RolloutTooSlowOrStuck alerts.

Weekly release

• Updates GMX to v1.0.1.
• Adds new alert for stuck DaemonSet rollouts.
• RateLimiter alert now tolerates 1 error before alerting.
• snmp-exporter now mounts its config from a ConfigMap, not a file copied into an emptyDir.