0.4.7 - Discard unhealthy nodes

Only consider consul nodes for a service passing all checks

0.4.6 - OpenSearch SM support

Add support for OpenSearch SM

0.4.5: Authenticate on kibana nodes (#12)

For Opensearch use case, we need to authenticate with username and password to probe the nodes.

Even if there is an authentication and the node does not need it, the node will ignore the authentication and answer the request

Add usage of a specific endpoint port (#11)

When pulling cluster from Consul, the port exposed is not necessarily the one that must be used in order to make cluster level calls.
This feature allows to set a different endpoint port to make these calls on the ElasticSearch cluster.

By default, the endpoint port is the one exposed in Consul..

0.4.3: disable consul agent caching during kibana node discovery

Full Changelog: 0.4.2...0.4.3

0.4.2: Merge pull request #10 from geobeau/master

Add build job

0.4.1: Use fqdn when possible

the svc.Node is not always fqdn, we want to use fqdn to ease intersection with other metrics

0.4.0

Add probing on restore snapshot feature (#9)

* Add probing on restore snapshot feature

Check that the durability index can well be restored from a specific snapshot policy
Only available with ES 7 as it rely on snapshot policy

0.3.1

Use discovered scheme for kibana probing (#8)

0.3.0

[BREAKING] - Rework probe (#6)

* Cat api return raw lines so can't be parsed at json

Also querying cat doesn't perform any search so changing metrics naming
to cat latency instead of search latency

* Migrate to kong command line parser

* Bump dependency

* Remove useless config file usage

* Move discovery of new node to a watcher

* Watcher now discover clusters instead of node

Kibana and ES have 2 different probes per cluster each being in charge of it's own set of
nodes to monitor
Watcher is only in charge of discovering/starting new probes for ES/Kibana clusters and remove them
Each probe cluster is then in charge of dicovering it's attached node and doing the appropriate monitoring
This will permits to add per cluster search and durability check

* Add cluster durability and latency probe

* Only use one consul client for all components

Currently each call on discovery leads to a consul client being initialised

* Few refacto

* Use different tick per ES probe

This permitts to run more often the latency test and to base it from a rate per min
instead running it only once every 30s

* Add latency histogram

* Few corrections

* Take in account remarks

* Update README.md