Search through HTCondor user log files to extract information, identify patterns, and build time-based graphs of resource utilization.
NOTE: HTCAnalyze depends on the htcondor
Python module, which currently
only runs on Linux.
This software can be installed directly from GitHub.
pip install git+https://github.com/psyinfra/HTCAnalyze.git
As a general recommendation, we encourage installing HTCAnalyze in a virtual environment. If you are not already familiar with Python virtual environments, we highly recommend that you read up on them.
A configuration file (htcanalyze.conf
) can be used to control the default
behavior. This is especially useful for system administrators who wish to tune
the warning and error thresholds to their site requirements.
It can also be passed directly as a flag.
htcanalyze -c htcanalyze.conf paths
See CONFIG for more information.
HTCAnalyze has two modes of operation:
analyze
: provides detailed output, including a RAM
histogram, HTCondor errors, and more.
This is the default if only one file is given.
summarize
: provides summarized output collected by each state of a job
(e.g. waiting, running, terminated, etc.)
This is the default if more than one file is given.
Example output:
analyze mode:
summarize mode:
htcanalyze -h
or
man htcanalyze (uses more detailed man page)
htcanalyze logs/job_5991_0.log
htcanalyze logs/
Tests can be found in the tests/
dir. To invoke the tests, run:
Dependencies for tests are listed in requirements-test.txt
and can be
installed by running:
pip install -U -r requirements-test.txt
To run the tests:
pytest tests
or with coverage:
pytest --cov=htcanalyze tests