Logster is a utility for reading log files and generating metrics in Graphite or Ganglia. It is ideal for visualizing trends of events that are occurring in your application/system/error logs. For example, you might use logster to graph the number of occurrences of HTTP response code that appears in your web server logs.
Logster maintains a cursor, via logtail, on each log file that it reads so that each successive execution only inspects new log entries. In other words, a 1 minute crontab entry for logster would allow you to generate near real-time trends in Graphite or Ganglia for anything you want to measure from your logs.
This tool is made up of a framework script, logster, and parsing scripts that are written to accommodate your specific log format. Two sample parsers are included in this distribution. The parser scripts essentially read a log file line by line, apply a regular expression to extract useful data from the lines you are interested in, and then aggregate that data into metrics that will be submitted to either Ganglia or Graphite. Take a look through the sample parsers, which should give you some idea of how to get started writing your own.
The logster project was created at Etsy as a fork of ganglia-logtailer (https://bitbucket.org/maplebed/ganglia-logtailer). We made the decision to fork ganglia-logtailer because we were removing daemon-mode from the original framework. We only make use of cron-mode, and supporting both cron- and daemon-modes makes for more work when creating parsing scripts. We care strongly about simplicity in writing parsing scripts -- which enables more of our engineers to write log parsers quickly.
Logster depends on the "logtail" utility that can be obtained from the logcheck package, either from a Debian package manager or from source:
RPMs for logcheck can be found here:
Once you have logtail installed via the logcheck package, you make want to look
over the actual logster script itself to adjust any paths necessary. Then the
only other thing you need to do is run the installation commands from the
$ sudo python setup.py install
You can test logster from the command line. There are two sample parsers: SampleLogster, which generates stats from an Apache access log; and Log4jLogster, which generates stats from a log4j log. The --dry-run option will allow you to see the metrics being generated on stdout rather than sending them to either Ganglia or Graphite.
$ sudo /usr/sbin/logster --dry-run --output=ganglia SampleLogster /var/log/httpd/access_log $ sudo /usr/sbin/logster --dry-run --output=graphite --graphite-host=graphite.example.com:2003 SampleLogster /var/log/httpd/access_log
Additional usage details can be found with the -h option:
$ ./logster -h usage: logster [options] parser logfile Tail a log file and filter each line to generate metrics that can be sent to common monitoring packages. Options: -h, --help show this help message and exit -p METRIC_PREFIX, --metric-prefix=METRIC_PREFIX Add prefix to all published metrics. This is for people that may multiple instances of same service on same host. -x METRIC_SUFFIX, --metric-suffix=METRIC_PREFIX Add suffix to all published metrics. This is for people that may multiple instances of same service on same host. --parser-help Print usage and options for the selected parser --parser-options=PARSER_OPTIONS Options to pass to the logster parser such as "-o VALUE --option2 VALUE". These are parser-specific and passed directly to the parser. --gmetric-options=GMETRIC_OPTIONS Options to pass to gmetric such as "-d 180 -c /etc/ganglia/gmond.conf" (default). These are passed directly to gmetric. --graphite-host=GRAPHITE_HOST Hostname and port for Graphite collector, e.g. graphite.example.com:2003 --aws-key=AWS_KEY Amazon credential key --aws-secret-key=AWS_SECRET_KEY Amazon credential secret key -s STATE_DIR, --state-dir=STATE_DIR Where to store the logtail state file. Default location /var/run -o OUTPUT, --output=OUTPUT Where to send metrics (can specify multiple times). Choices are 'graphite', 'ganglia', or 'stdout'. -d, --dry-run Parse the log file but send stats to standard output. -D, --debug Provide more verbose logging for debugging.