elasticsearch-cat

http://chasemp.github.io/2015/08/05/exploring-elasticsearch-_cat/

==== Preamble

If you want to index and search your documents with lucene then
Elasticsearch seems to be the popular way to do so while adding
redundandcy, clustering, and an API on top of Lucene. Elasticsearch
has a collection of REST API's with one subset not returning json at all.
The _cat API is meant for humans. While there are CLI tools made
for parsing json most of the popular tools we are used to are more
natively served with raw textual data.

Making sure the API is responding

It's worth noting that when a cluster is new and has no designated
master nodes this API seems to hang indefinitely. If there is a splitbrain
problem the same thing occurs. This can be really concerning if you are
in trouble master role wise as the API you expect to query will not respond.

[2015-09-16 17:04:08,233][DEBUG][action.admin.cluster.health] \
[elastic-node] no known master node, scheduling a retry
There can be duplicate information across API calls. For instance, in order
to make sense of the shards you undoubtably need the index status. Whatever
endpoint is more appropriate is ultimately a matter of perspective.

curl -s 'http://localhost:9200/'
{
“status” : 200,
“name” : “elastic-node”,
“cluster_name” : “my-cluster-name”,
“version” : {
“number” : “1.7.1”,
“build_hash” : “”,
“build_timestamp” : “2015-07-29T09:54:16Z”,
“build_snapshot” : false,
“lucene_version” : “4.10.4”
},
“tagline” : “You Know, for Search”
}

Cat

curl -s localhost:9200/_cat/
aliases
allocation
count - document count
fielddata - field data by index
health - cluster health
indices - index overview and status
master - find current master node
nodes - node status(es)
pending_tasks - task backlog information 
plugins - plugins and versions across all nodes
recovery - index/shard status with movement info
thread pool - tons of info on active threads across nodes
shards - tons of info on shard states
segments - low level segment info

_cat tricks

Seeing all possible columns

curl 'localhost:9200/_cat/master?help'
id | | node id
host | h | host name
ip | | ip address
node | n | node name

Column headers

A note is this will not show columns for which there are no data. This is
noticeable for a few endpoints such as fielddata especially.

curl 'localhost:9200/_cat/master?v'
id host ip node
1234 elastic-node 10.0.0.143 elastic-node

Picking relevant column headers

curl localhost:9200/_cat/master?h='host'

elastic-node

curl localhost:9200/_cat/master?h='host,ip'

elastic-node 10.0.0.143
Watching shard movement

curl -s 'localhost:9200/_cat/recovery?active_only=true' | \
awk '{printf "%-19s %6s %-12s => %-12s\n", $1, $13, $6, $7}' | \
sort -n -k2 -r

myindex  59.1% elastic-node1  => elastic-node2
Reference

https://aphyr.com/posts/323-call-me-maybe-elasticsearch-1-5-0