Implement Burrow input plugin

Godeps

@@ -17,7 +17,6 @@ github.com/docker/go-connections 990a1a1a70b0da4c4cb70e117971a4f0babfbf1a
 github.com/eapache/go-resiliency b86b1ec0dd4209a588dc1285cdd471e73525c0b3
 github.com/eapache/go-xerial-snappy bb955e01b9346ac19dc29eb16586c90ded99a98c
 github.com/eapache/queue 44cc805cf13205b55f69e14bcb69867d1ae92f98
-github.com/fortytw2/leaktest 3b724c3d7b8729a35bf4e577f71653aec6e53513
 github.com/eclipse/paho.mqtt.golang aff15770515e3c57fc6109da73d42b0d46f7f483
 github.com/go-logfmt/logfmt 390ab7935ee28ec6b286364bba9b4dd6410cb3d5
 github.com/go-sql-driver/mysql 2e00b5cd70399450106cec6431c2e2ce3cae5034

plugins/inputs/burrow/README.md

@@ -1,102 +1,95 @@
 # Telegraf Plugin: Burrow
 
-Collect Kafka topics and consumers status
-from [Burrow](https://github.com/linkedin/Burrow) HTTP API.
+Collect Kafka topic, consumer and partition status
+via [Burrow](https://github.com/linkedin/Burrow) HTTP [API](https://github.com/linkedin/Burrow/wiki/HTTP-Endpoint).
 
-### Configuration:
+Supported Burrow version: `1.x`
+
+### Configuration
 
 ```
-[[inputs.burrow]]
-  ## Burrow endpoints in format "sheme://[user:password@]host:port"
-  ## e.g.
-  ##   servers = ["http://localhost:8080"]
-  ##   servers = ["https://example.com:8000"]
-  ##   servers = ["http://user:[email protected]:8000"]
-  ##
-  servers = [ "http://127.0.0.1:8000" ]
-
-  ## Prefix all HTTP API requests.
-  #api_prefix = "/v2/kafka"
+  ## Burrow API endpoints in format "schema://host:port".
+  ## Default is "http://localhost:8000".
+  servers = ["http://localhost:8000"]
 
-  ## Maximum time to receive response.
-  #timeout = "5s"
+  ## Override Burrow API prefix.
+  ## Useful when Burrow is behind reverse-proxy.
+  # api_prefix = "/v3/kafka"
 
-  ## Optional, gather info only about specific clusters.
-  ## Default is gather all.
-  #clusters = ["clustername1"]
+  ## Maximum time to receive response.
+  # response_timeout = "5s"
 
-  ## Optional, gather stats only about specific groups.
-  ## Default is gather all.
-  #groups = ["group1"]
+  ## Limit per-server concurrent connections.
+  ## Useful in case of large number of topics or consumer groups.
+  # concurrent_connections = 20
 
-  ## Optional, gather info only about specific topics.
-  ## Default is gather all
-  #topics = ["topicA"]
+  ## Filter out clusters by providing list of glob patterns.
+  ## Default is no filtering.
+  # clusters = []
 
-  ## Concurrent connections limit (per server), default is 4.
-  #max_concurrent_connections = 10
+  ## Filter out consumer groups by providing list of glob patterns.
+  ## Default is no filtering.
+  # groups = []
 
-  ## Internal working queue adjustments (per measurement, per server), default is 4.
-  #worker_queue_length = 5
+  ## Filter out topics by providing list of glob patterns.
+  ## Default is no filtering.
+  # topics = []
 
   ## Credentials for basic HTTP authentication.
-  #username = ""
-  #password = ""
+  # username = ""
+  # password = ""
 
   ## Optional SSL config
-  #ssl_ca = "/etc/telegraf/ca.pem"
-  #ssl_cert = "/etc/telegraf/cert.pem"
-  #ssl_key = "/etc/telegraf/key.pem"
-
-  ## Use SSL but skip chain & host verification
-  #insecure_skip_verify = false
+  # ssl_ca = "/etc/telegraf/ca.pem"
+  # ssl_cert = "/etc/telegraf/cert.pem"
+  # ssl_key = "/etc/telegraf/key.pem"
+  # insecure_skip_verify = false
 ```
 
-Due to the nature of Burrow API (REST), each topic or consumer metric
-collection requires an HTTP request, so, in order to keep things running
-smooth, consider two parameters:
+### Partition Status mappings
 
-1. `max_concurrent_connection` - limit maximum number of concurrent HTTP
-requests (per server).
-2. `worker_queue_length` - number of concurrent workers processing
-each measurement (per measurement, per server).
+* `OK` = 1
+* `NOT_FOUND` = 2
+* `WARN` = 3
+* `ERR` = 4
+* `STOP` = 5
+* `STALL` = 6
 
-Just keep in mind, each worker in queue requires an HTTP connection,
-so keep `max_concurrent_connection` and `worker_queue_length` balanced
-in ratio 2:1.
+> unknown value will be mapped to 0
 
-### Partition Status mappings
+### Fields
 
-* OK = 1
-* NOT_FOUND = 2
-* WARN = 3
-* ERR = 4
-* STOP = 5
-* STALL = 6
+* `burrow_group` (one event per each consumer group)
+  - status (string, see Partition Status mappings)
+  - status_code (int, `1..6`, see Partition status mappings)
+  - parition_count (int, `number of partitions`)
+  - total_lag (int64, `totallag`)
+  - lag (int64, `maxlag.current_lag || 0`)
 
-### Measurements & Fields:
+* `burrow_partition` (one event per each topic partition)
+  - status (string, see Partition Status mappings)
+  - status_code (int, `1..6`, see Partition status mappings)
+  - lag (int64, `current_lag || 0`)
+  - offset (int64, `end.timestamp`)
+  - timestamp (int64, `end.timestamp`)
 
-- burrow_topic (one event for each topic offset)
+* `burrow_topic` (one event per topic offset)
   - offset (int64)
 
-- burrow_consumer
-  - start.offset (int64)
-  - start.lag (int64)
-  - start.timestamp (int64)
-  - end.offset (int64)
-  - end.lag (int64)
-  - end.timestamp (int64)
-  - status (1..6, see Partition status mappings)
 
 ### Tags
 
-- burrow_topic
-  - cluster
-  - topic
-  - partition
+* `burrow_group`
+  - cluster (string)
+  - group (string)
+
+* `burrow_partition`
+  - cluster (string)
+  - group (string)
+  - topic (string)
+  - partition (int)
 
-- burrow_consumer
-  - cluster
-  - group
-  - topic
-  - partition
+* `burrow_topic`
+  - cluster (string)
+  - topic (string)
+  - partition (int)