Add logging documentation, including vipin-created logs (#199)

- *Category*:  documentation
- *JIRA issue*: [MIC-4757](https://jira.ihme.washington.edu/browse/MIC-4757)

Changes and notes
- Adds a logging page to the docs, describes all the logs in general, with sections for each "area" of logs.

.zenodo.json

@@ -39,5 +39,5 @@
     }
   ],
   "access_right": "open",
-  "description": "Archival release of Vivarium Cluster Tools, a Python package that makes running ``vivarium`` simulations at scale on a Univa Grid Engine cluster easy."
+  "description": "Archival release of Vivarium Cluster Tools, a Python package that makes running ``vivarium`` simulations at scale on a Slurm cluster easy."
 }

CHANGELOG.rst

@@ -1,3 +1,7 @@
+**1.5.1 - 12/15/23**
+
+ - Add logging documentation for psimulate
+
 **1.5.0 - 10/27/23**
 
  - Remove default results directory for 'psimulate run'

README.rst

@@ -13,7 +13,7 @@ Vivarium Cluster Tools
     :alt: Documentation Status
 
 Vivarium cluster tools is a python package that makes running ``vivarium``
-simulations at scale on a Univa Grid Engine cluster easy.
+simulations at scale on a Slurm cluster easy.
 
 Installation
 ------------

docs/source/index.rst

@@ -7,7 +7,7 @@
 Vivarium Cluster Tools Documentation
 ====================================
 Vivarium cluster tools is a python package that makes running ``vivarium``
-simulations at scale on a Univa Grid Engine cluster easy.
+simulations at scale on a Slurm cluster easy.
 
 .. toctree::
    :maxdepth: 2
@@ -16,5 +16,6 @@ simulations at scale on a Univa Grid Engine cluster easy.
    distributed_runner
    yaml_basics
    branch
+   logging
    api_reference/index
    glossary

docs/source/logging.rst

@@ -0,0 +1,43 @@
+.. toctree::
+   :maxdepth: 2
+   :caption: Contents:
+
+Logging
+============
+
+Sometimes, even with perfect code, things can go wrong at sufficient scale.
+When they do, it's useful to look to the logs to see what happened.  ``psimulate``
+logs to the results directory, in a subdirectory called ``logs``. Inside that directory,
+there will be a directory for each simulation run or restart. If neither
+``psimulate restart`` nor ``psimulate expand`` was
+ever used for the run, there will be only one directory for the run.
+
+Top-level logs
+----------------
+At the top-level of the directory, there will be text and JSON-formatted main log files.
+These are the log files for the runner process. There will also be a log file for each
+Redis database process, which will be named ``redis.p<port>.log``. Per-worker logs are
+in ``cluster_logs`` and ``worker_logs`` directories, described below.
+
+Cluster logs
+-------------
+The ``cluster_logs`` directory contains logs from the the array job processes. Each worker job
+has its own file. The contents of these are similar to what you will find in the ``worker_logs``
+directory, but a superset. The logs in the ``cluster_logs`` directory contain information about Redis
+heartbeats and other cluster-related information.
+
+Worker logs
+-------------
+The ``worker_logs`` directory contains logs from the the worker processes as they relate
+running simulations. Additionally this directory contains performance logs that
+are described in the next section.
+
+Performance logs
+-----------------
+As part of the VIPIN (VIvarium Performance INformation) feature, ``psimulate`` gathers
+per-worker performance information. This information is summarized at the end of the parallel
+runs and stored in the ``worker_logs`` directory as ``log_summary.csv``. This file
+contains metadata identifying the run and the worker host, execution timing information, and CPU,
+disk, and network performance counters. The intent of this logging is to allow users to understand the
+performance characteristics of their simulations and in the event of suspicious performance,
+to be able to correlate outlier performance characteristics to cluster and hardware events.