This is a cloned repo from https://github.com/zooldk/couchbase_collector
It expands the previous repo by collecting all metrics under:
"nodes" : [ "interestingStats" ]
"basicStats"
"quota"
keys from the returned couchbase json object.
You need to place it under a directory called couchbase_collector under /usr/share/diamond/collectors or wherever your diamond collector code resides
You also need a configuration file CouchBaseCollector.conf under /etc/diamond/collector (or wherever you collector configuration files reside) with the following content:
enabled = True
That should be it.
Configuration parameters for couchbase (host, port, path, name, password) need to be placed inside the get_default_config method in the CouchBaseCollector class.
Once you restart diamond you should be able to see in the log file entries like (metrics are from the beer-sample example in couchbase):
mytestserver.beer-sample.basicStats.diskUsed: 25565808
mytestserver.beer-sample.basicStats.memUsed: 35407008
mytestserver.beer-sample.basicStats.diskFetches: 0.0
mytestserver.beer-sample.basicStats.quotaPercentUsed: 33.7667541504
mytestserver.beer-sample.basicStats.opsPerSec: 0.0
mytestserver.beer-sample.basicStats.dataUsed: 24288256
mytestserver.beer-sample.basicStats.itemCount: 7303
mytestserver.beer-sample.nodes.couch_views_data_size: 543142
mytestserver.beer-sample.nodes.couch_docs_actual_disk_size: 56489751
mytestserver.beer-sample.nodes.couch_views_actual_disk_size: 564012
mytestserver.beer-sample.nodes.ops: 0.0
mytestserver.beer-sample.nodes.vb_replica_curr_items: 0
mytestserver.beer-sample.nodes.cmd_get: 0.0
mytestserver.beer-sample.nodes.mem_used: 99496312
mytestserver.beer-sample.nodes.get_hits: 0.0
mytestserver.beer-sample.nodes.curr_items: 7889
mytestserver.beer-sample.nodes.ep_bg_fetched: 0.0
mytestserver.beer-sample.nodes.curr_items_tot: 7889
mytestserver.beer-sample.nodes.couch_docs_data_size: 55645339
mytestserver.beer-sample.quota.ram: 104857600
mytestserver.beer-sample.quota.rawRAM: 104857600
In clustered environments the nodes array holds "interestingStats" for all cluster members. The Collector will only ship interestingStats referring to the host being queried.
Each cluster member is assumed to run a separate instance of diamond (no point running a couchbase cluster one one machine) so interestingStats of each will appear under the corresponding host.