support-script-analyzer

A set of scripts to extract information from archives generated by the enterprise support script. The output of each job is in json.

Requirements

• jq
• GNU awk
• bash 4+
• GNU coreutils

Usage

support_parse.sh <job> <support_archive.tar.gz>

Currently implemented jobs are

• db_sizes
• modules
• tech_check

json output can be converted to HTML with some basic styling by saving to a file and passing the filename as an argument to convert.py. The required Python libraries can be installed with pip install -r requirements.txt.

sar2influxdb

Tools for importing Puppet metrics into time series databases

Usage:

sar2influxdb.rb [options] [sadd|sadd.gz] [...]

--pattern PATTERN Glob pattern of files to load.

• Must be provided if no files are passed.

--db-host HOSTNAME|IP_ADDRESS

• Hostname to submit converted data to.

• Leave blank to print converted data to stdout.

--db-name NAME

• Database name to submit converted data to.

• Required if --db-host is used.

-h, --help

• Show help

--debug

• Enable backtraces from errors.

--version

• Show version

Limitations

HTTP output not fully plumbed in. Needs to be done in batches of ~10,000 lines as there is a limit to how much data InfluxDB will accept in a single POST. For now, dump stdout to a file, split -l10000 it, then use curl to POST the resulting x* files with --data-binary.

This tool locates the sadf executable, which is used to extract and format data stored in SAR archives. This tool is commonly provided by the sysstat package and version 11.1.1 or newer is required.

You may have to run this script in an environment that has the same version of sysstat as the environment that produced the archive. Docker containers are a great way to do this.