[coordinator] Add support for prometheus remote write compatible backend

[Antanukas]

What this PR does / why we need it:

Adds support for new backend type - prometheus remote write. This enables using M3 Coordinator and M3 Aggregator with Prometheus Remote Write compatible API.

Special notes for your reviewer:

Read part is not implemented mainly because using remote read API is not efficient since it requires to load raw timeseries.

In this PR integration with aggregator is not fully implemented. For now one can specify endpoints with no storage policy only so they will receive raw writes.

I plan to fix this in subsequent PR.

Example of configuration when data aggregation will be available:

backend: prom-remote

prometheusRemoteBackend:
  endpoints:
    - name: raw
      address: "http://raw-metrics/api/v1/prom/remote/write"
    - name: aggregated_30d
      address: "http://aggregated_30d/api/v1/prom/remote/write"
      storagePolicy:
        retention: 720h
        resolution: 1m
    - name: aggregated_60d
      address: "http://aggregated_60d/api/v1/prom/remote/write"
      storagePolicy:
        retention: 1440h
        resolution: 1m
		
clusterManagement:
  etcd:
    env: test-env
    zone: embedded
    service: m3db
    cacheDir: /var/lib/m3kv
    etcdClusters:
    - zone: embedded
      endpoints:
      - http://etcd-0:2379
      - http://etcd-1:2379
      - http://etcd-2:2379

Does this PR introduce a user-facing and/or backwards incompatible change?:

This feature will be available and can be enabled using configuration switch. 

Does this PR require updating code package or user-facing documentation?:

Documentation will follow after this PR is merged.

[codecov]

Codecov Report

Merging #3742 (4a18a72) into master (9fea52c) will decrease coverage by 0.0%.
The diff coverage is n/a.

@@           Coverage Diff            @@
##           master   #3742     +/-   ##
========================================
- Coverage    57.0%   57.0%   -0.1%     
========================================
  Files         552     552             
  Lines       62950   62950             
========================================
- Hits        35919   35913      -6     
- Misses      23840   23841      +1     
- Partials     3191    3196      +5     

Flag	Coverage Δ
aggregator	63.3% <ø> (ø)
cluster	∅ <ø> (∅)
collector	58.4% <ø> (ø)
dbnode	60.7% <ø> (-0.1%)	⬇️
m3em	46.4% <ø> (ø)
metrics	19.7% <ø> (ø)
msg	74.3% <ø> (-0.2%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9fea52c...4a18a72. Read the comment docs.

[vpranckaitis]

A few initial comments, will continue reviewing tomorrow.

[vpranckaitis]

LGTM, though it would be best if someone else would take a look, too.

[linasm]

So you can return a generic error now, and TODO just logging.

[linasm]

LGTM with a few comments.

[robskillington]

LGTM except (outside of nits/minor comments) I do feel relatively strongly that we return an error when retention and resolution doesn't match any of the endpoints listed before we merge this (referencing this comment here https://github.com/m3db/m3/pull/3742/files#r713168023)