replace firehose_exporter, move to attic legacy job and dependencies

[psycofdj]

This PR supersedes #414 in bumping firehose_exporter to its new major release v7.0.0.

This major release has a non-backward compatible impact on generated metrics and, therefore, on associated jobs cloudfoundry_dashboards and cloudfoundry_alerts.

In order to allow users to stick to legacy implementation and dashboards, previous jobs and packages has been kept and renamed with -attic suffix:

• packages/firehose_exporter-attic
• jobs/firehose_exporter-attic
• jobs/cloudfoundry_dashboards-attic
• jobs/cloudfoundry_alerts-attic

These jobs and packages won't benefit from further maintenance.

Additionally, new ops-files are provided to easily switch back to legacy implementation:

• manifests/operators/deprecated/monitor-cf-attic.yml, to use in addition of manifests/operators/monitor-cf.yml
• manifests/operators/deprecated/enable-cf-loggregator-v2.yml which replaces deleted manifests/operators/enable-cf-loggregator-v2.yml
• manifests/operators/deprecated/cf/add-prometheus-uaa-clients-attic.yml to use in addition of manifests/operators/cf/add-prometheus-uaa-clients.yml in cloud foundry deployment

Finally, the mess between "v2" and "original" dashboards in cloudfoundry_dashboards has been cleaned up:

• v2 dashboards now replace the original ones
• they have been renamed to match original names
• v2 tags has been removed and associated dashboard list dropdown has been removed.

Using the manifests/operators/deprecated/monitor-cf-attic.yml ops-file will also revert this change.

Note: new firehose blob firehose_exporter/firehose_exporter_7.0.0.linux-amd64.tar.gz has already been uploaded to AWS S3

[benjaminguttmann-avtq]

I think the README.md is missing some adjustments for the new firehose, right? Can you please check that https://github.com/bosh-prometheus/prometheus-boshrelease#monitoring-cloud-foundry

[psycofdj]

I think the README.md is missing some adjustments for the new firehose, right? Can you please check that https://github.com/bosh-prometheus/prometheus-boshrelease#monitoring-cloud-foundry

sure, you are right

[psycofdj]

here you go @benjaminguttmann-avtq, see d8ad880

[benjaminguttmann-avtq]

I have another question: https://github.com/bosh-prometheus/prometheus-boshrelease/blob/6e757899bdf36fd53eb9d16917299d52e7656656/manifests/operators/monitor-cf.yml#L67

what happens if you use that exactly like it is in there? Will that cause issues because certificates do not match the ones from loggregrator?

[psycofdj]

Prometheus-boshrelease users have to specify a value for this variable. I will add details in the README.md in ad additional commit .

The idea here is to generate a authentication certificate. In order to be accepted by reverse_log_proxy (lpr), this certificate has to be generated from the loggregator CA.
This CA is already in credhub but belongs to cloud foundry deployment (generated by cf-deployment). To access it from another deployment, we must specify its full path, something like /bosh-mydirector/cf/loggregator_ca.

[psycofdj]

here you are: 2d354e5