esmc is a wrapper for elasticsearch.

The primary function of esmc is to provide support for interacting with multiple elasticsearch clusters, and secondarily to provide per-request instrumentation (via prometheus).

mc := esmc.MustNewMultiCluster("es://:9200?name=one&mode=on es://:9201?name=two&mode=dark")

// Can be used as an es.Searcher or es.MultiSearcher
mc.Search(&es.SearchRequest{
	Params: es.SearchParams{
		Indices: []string{"twitter"},
		Types:   []string{"tweet"},
	},
	Query:   es.MatchAllQuery(),
})

// Can be iterated to access invidivual clusters, which implement the full
// es.Cluster API.
for _, c := range mc {
	c.Index(&es.IndexRequest{
		es.IndexParams{Index: "twitter", Type: "tweet", Id: "1"},
		map[string]interface{}{},
	})
}

esmc provides support for multiple elasticsearch clusters, with the goal of simplifying failover and testing.

The default behavior for Search and MultiSearch is the following:

• the request is sent to all clusters in "on" or "dark" mode.
• the first response from an "on" cluster is returned to the user.
• all other responses are discarded (after reporting metrics).

The following metrics are exposed through the prometheus client's registry:

• esmc_requests: A counter of the total number of requests to an ES cluster
• esmc_request_total_duration_nanoseconds: The total amount of time spent executing requests (nanoseconds)
• esmc_request_durations_nanoseconds: The amounts of time spent executing requests (nanoseconds)
• esmc_reported_request_total_duration_nanoseconds: The total amount of time spent executing requests as reported by elasticsearch (nanoseconds)
• esmc_reported_request_durations_nanoseconds: The amounts of time spent executing requests as reported by elasticsearch (nanoseconds)

Each metric is labelled with:

• cluster: the cluster name (from the cluster's config)
• cluster_mode: the cluster's mode (on, off, dark)
• request_type: search, multi_search, index, create, update, delete, bulk, execute
• outcome: success, failure