- Jobs can now be inserted and queried with arbitrary domain IDs
- Collect agent uses domain ID accordingly during inserts

- Now using an optional "domain ID" as Cassandra partition key
- Can correspond to HPC system, SLURM partition, etc.
- Using materialized views to optimize performance on range queries

- Net-SNMP's persistent directory path can be now specified as a variable

- Reduced code duplication
- Handling some uncommon error conditions
- Preventing error tangling
- Adding methods to fetch pending and finished jobs

Use getPublicSensorNames() to retrieve lit of sensors for wildcard matching instead of getPublicSensorsVerbose()

- If the tool is launched within a job step, the job id will be updated
accordingly (e.g., 12345.1)
- If the SLURM_JOB_USER variable is not defined, we use USER
- Added node substitution for the DEEP-EST prototype

- Operations based on sensors with custom scaling factors would output
wrong results

- Job Operator job queries now use the timestamp for the next scheduled
invocation to define their time range
- Additional timestamp checks in PersystSQL

- Added a "raw" mode in which only the average is used as feature for
each sensor. Useful when users wish to build signatures with a different
plugin and then pipeline them into the model;
- Restored the "targetDistance" parameter for the Classifier plugin;
- Fixed a minor bug affecting the number of samples used for training.

- Fixed a memory leak
- Improved exception management
- General refactoring of the code

- Smoothing is not hardcoded anymore and the server picks automatically
the best smoother based on the available sensor operations
- Requires a "smootherRegex" configuration parameter to identify the
smoothing operations