Run queries against Kibana's Elasticsearch that gets logs from Logstash. Forked from Wikia's kibana.py.

pip install elasticsearch-query

from elasticsearch_query import ElasticsearchQuery
es_query = ElasticsearchQuery(es_host='es.prod', since=12345, period=900, index_prefix='logstash-my-app')

es_host needs to be specified with a host of Elasticsearch instance to connect.

Provide either since (absolute timestamp) or period (last N seconds):

• since: UNIX timestamp data should be fetched since (if None, then period specifies the last n seconds).
• period: period (in seconds) before now() to be used when since is empty (defaults to last 15 minutes).

index_prefix argument will be used to build indices names to query in. They should follow the index-name-YYYY.MM.DD naming convention, e.g. logstash-my-app-2014.08.19.

Returns data matching the given query (provided as a dict).

es_query.get_rows(match={"tags": 'edge-cache-requestmessage'}, limit=2000)

• match: query to be run against log messages (ex. {"@message": "Foo Bar DB queries"}).
• limit: the number of results (defaults to 10).

Returns data matching the given query string (provided as a Lucene query).

es_query.query_by_string(query='@message:"^PHP Fatal"', limit=2000)
es_query.query_by_string(query='@message:"^PHP Fatal"', fields=['@message', '@es_query_host'], limit=2000)

• query: query string to be run against log messages (ex. @message:"^PHP Fatal").
• fields: optional list of fields to fetch
• limit: the number of results (defaults to 10).

Returns data matching the given SQL query.

This feature requires non-OSS version of Elasticsearch ("To be clear, while the X-Pack source code is now available in the public repositories, it isn’t under an Open Source license").

es_query.query_by_sql(sql='SELECT host FROM "app-requests" WHERE host = \'app2.prod\'')

• sql: SQL query to be run

Returns number of matching entries

es_query.count(query='@message:"^PHP Fatal"')

elasticsearch-query comes with integration tests suite. .travis.yml will install elasticsearch OSS version and run them.

But you can also run it locally. Simply pass ES_TEST_HOST=<elasticsearch IP> env variable when running tests (make test).