CLI tool to pull statistics from Elasticsearch.
The tool uses docopt to describe command line language and supports the following options:
Usage:
watches cluster_health [-i=INTERVAL -d=DURATION --url=URL -bltsv -f=FILTER...] [(--cacert=CACERT --cert=CERT --key=KEY) | (--cacert=CACERT)] [(--username=USERNAME --password=PASSWORD)] [--header=HEADER...] [--level=LEVEL --local]
watches cluster_state [-i=INTERVAL -d=DURATION --url=URL -bltsv -f=FILTER...] [(--cacert=CACERT --cert=CERT --key=KEY) | (--cacert=CACERT)] [(--username=USERNAME --password=PASSWORD)] [--header=HEADER...] [--local --index=INDEX --metric=METRIC]
watches cluster_stats [-i=INTERVAL -d=DURATION --url=URL -bltsv -f=FILTER...] [(--cacert=CACERT --cert=CERT --key=KEY) | (--cacert=CACERT)] [(--username=USERNAME --password=PASSWORD)] [--header=HEADER...]
watches nodes_stats [-i=INTERVAL -d=DURATION --url=URL -bltsv -f=FILTER...] [(--cacert=CACERT --cert=CERT --key=KEY) | (--cacert=CACERT)] [(--username=USERNAME --password=PASSWORD)] [--header=HEADER...] [--metric=METRIC]
watches nodes_info [-i=INTERVAL -d=DURATION --url=URL -bltsv -f=FILTER...] [(--cacert=CACERT --cert=CERT --key=KEY) | (--cacert=CACERT)] [(--username=USERNAME --password=PASSWORD)] [--header=HEADER...] [--node_id=NODE_ID --metric=METRIC]
watches indices_stats [-i=INTERVAL -d=DURATION --url=URL -bltsv -f=FILTER...] [(--cacert=CACERT --cert=CERT --key=KEY) | (--cacert=CACERT)] [(--username=USERNAME --password=PASSWORD)] [--header=HEADER...] [--level=LEVEL --index=INDEX]
watches nodes_hotthreads [-i=INTERVAL -d=DURATION --url=URL -bsv] [(--cacert=CACERT --cert=CERT --key=KEY) | (--cacert=CACERT)] [(--username=USERNAME --password=PASSWORD)] [--header=HEADER...] [--node_id=NODE_ID --threads=THREADS --delay=DELAY --type=TYPE]
watches -h
watches --version
Options:
-d=DURATION, --duration=DURATION How long the watches should run in seconds. Use value '-1' to run forever. [default: 0].
-i=INTERVAL, --interval=INTERVAL Interval between data retrievals. Apply if 'duration' > 0. [default: 3].
--url=URL URL of ES node HTTP endpoint [default: http://localhost:9200].
-b Disable python output buffering (not recommended for production use).
-l Single line output (no pretty-print JSON formatting).
-t, --timestamp Add timestamp field to data. The value is local datetime converted to UTC in ISO 8601 format.
-s, --sniff Turn on sniffing.
-v, --verbose Print more debug info: input options, ... etc.
-f=FILTER, --filter_path=FILTER Filter returned JSON (see http://elasticsearch-py.readthedocs.io/en/master/api.html#response-filtering)
--username=USERNAME Username to authenticate with
--password=PASSWORD Password to authenticate with
--cacert=CACERT Path to Certification Authority Certificate pem file
--cert=CERT Path to Client Certificate pem file
--key=KEY Path to Client Key pem file
--level=LEVEL Aggregation level of returned data, valid options: cluster, indices or shards [default: cluster].
--local Return the local node information instead of master node [default: false].
--node_id=NODE_ID A comma-separated list of node IDs or names to limit the returned information; use `_local` to return information from local node you're connecting to [default: ].
--index=INDEX A comma-separated list of index names; use `_all` or empty string to perform the operation on all indices.
--metric=METRIC A comma-separated list of metric names; use `_all` or empty string to perform the operation for all metrics.
--threads=THREADS Specify the number of threads to provide information for.
--delay=DELAY Delay (interval) to do the second sampling of threads.
--type=TYPE The type of threads to sample (default: cpu), valid choices are: 'cpu', 'wait', 'block'.
-h, --help Show this screen.
--version Show version.
--header=HEADER Custom HTTP header to add to the request (e.g. --header="X-Proxy-Remote-User: username")
Examples:
# Get cluster health from specified HTTP endpoint with added "timestamp" field in the response
$ watches cluster_health --timestamp --url=http://127.0.0.1:9200
# Get cluster health every 1 second, run forever until process is terminated
$ watches cluster_health --interval=1 --duration=-1
# Get cluster health every 1 second during next 10 seconds and use sniffing
$ watches cluster_health --interval=1 --duration=10 --sniff
# Alternatively, using short option notation
$ watches cluster_health -i 1 -d 10 -s
# Get cluster health from secured node
$ watches cluster_health \
--url=https://localhots:9200 \
--cacert /tmp/search-guard-ssl/example-pki-scripts/ca/chain-ca.pem \
--cert /tmp/search-guard-ssl/example-pki-scripts/kirk.crt.pem \
--key /tmp/search-guard-ssl/example-pki-scripts/kirk.key.pem
# Filter cluster health for status fields
$ watches cluster_health --level=shards -f status -f indices.*.status -f indices.*.shards.*.status
To connect to Elasticsearch cluster watches uses official
elasticsearch-py client which
can use Sniffing.
It is recommended to use --sniff option (see above) to enable sniffing for long running tasks.
If you've cloned this project, and want to install the library (and all development dependencies), the command you'll want to run is:
$ pip install -e .[test]
If you'd like to run all tests for this project (assuming you've written some), you would run the following command:
$ python setup.py test
This will trigger py.test, along with its popular coverage plugin.
Read Testing.md to learn more details.
Lastly, if you'd like to cut a new release of this CLI tool, and publish it to the Python Package Index (PyPI), you can do so by running:
$ python setup.py sdist bdist_wheel $ twine upload dist/*
This will build both a source tarball of your CLI tool, as well as a newer wheel build (and this will, by default, run on all platforms).
The twine upload command (which requires you to install the twine tool) will then securely upload your
new package to PyPI so everyone in the world can use it!
Built on top of skele-cli skeleton, read skele-cli blog post to learn more.
Watches CLI is licensed under the Apache License, Version 2.0.