nOutput_HTTP and nOutput_JSON should audit access as nOutput_Channels.

As done for nInput_RAIDMemory on 31225b8, nInput should implement a generic function (among others) to handle the extra for any input that requires it.

nOutput_EmailWarnings add max description to limit the size on emails of the description

If a plug has a timeInterval and also a cron entry the thread stuck detection mechanism only considers the timeInterval and will automatically restart nAttrMon if the expression on the cron provides a large time interval.

Be able to produce a backup file with the specific configuration (all active and inactive plugs). This could even go further to commit to SVN/GIT directly.

Enable the use of an extra description field for a more verbose description of the objective of each plug plus convert the attribute description javascript input plug (that is no longer working properly) into a reusable object that adds description to the defined objects upon attributes channel subscribe trigger.

Due to issue #64 a better thread stuck detection mechanism should be implemented for cron based plugs.

Option to redirect the log to stdout (without creating the log folder and rotating files). This would be useful for docker containers.

In some inputs the property "name" is also being added to attribute values. Doesn't seem to affect all attributes (for example semaphores are okay). Not yet sure if it's related with configuration or base code.
It's not a critical problem but rather annoying.

On nInput_Semaphores and nInput_CBPMSemaphores if a semaphore name has one or more "/" it will be interpreted as a sub-folder/sub-category. The names should be replaced by another character like "_".

Some Outputs and Validations could subscribe the Attribute values and Warnings channels. This way they wouldn't need a timeInterval or cron to execute and would be event driven. On the other hand this might only be the intended behaviour in some cases (notably for warnings).

When building nValidation Generic check expressions (e.g. expr) when an attribute has '\n' it causes the validation to fail with "InternalError: unterminated string literal".

Some nAttrMon implementations are "mirroring" the attribute values and warnings channels in ElasticSearch thus a generic nOutput_ElasticSearch is in order.

The current implementation uses a time based rotating housekeeping. But some attributes are best suited for a number of events based rotating housekeeping. Plus it should be possible to configure, per attribute, specific retention periods based on number of events of time.

p.s. This should apply also to nOutput_Oracle

Add more stats info to plugs like lastExecutedElapsedTime and similar.

Change the object to recieve a parameter called raidVersion string that will be default to 8. When set to 7 the exec command should call the old method to retrieve the schedulers for RAID 7.

Enable to use a definition file (yaml?) to select which warnings levels should be sent on the emails output.

A new option could be added to input plugs to merge the attribute with the existing attribute value by a key.

For example: the attribute "Audit DBs" could be the result of merging "Audit A", "Audit B" and "Audit C".
If each audit is a map composed of, for example:

[{
   "DB": "B",
   "Username": "xpto1",
   "LastLogin": "2018-01-01T01:02:03.123Z"
},
{
   "DB": "B",
   "Username": "xpto2",
   "LastLogin": "2018-01-01T01:05:03.123Z"
}
]

The end attribute "Audit DBs" has all the merged arrays of Audit A, B and C. Each time the individual audit input plug executes it replaces all entries from the provided key (with channel getSet).

There are some plugs, notably Output plugs, that are trigger by external events (like HTTP queries). So there isn't a need to update the internal plug info at a constant rate but rather they should query the internal channels for the requested info. Therefore some plugs could initialize but don't need to be executed every x time.

Using the new warnings on channels there are cases where closed warnings are not deleted for a long period of time.

Each attribute value is sent to ES like:

{
   "name_of_attribute": {
      "key1": "value1",
      "property2": "value2",
      "property3": "value3"
   },
   "stamp_key": "value_key"
}

but to filter by key1, that might be used in other attributes, several specific filters might be needed. If an execArg is added to provide a list of key field indexes (in this case [ "key1" ]) the resulting value sent to ES would be:

{
   "name_of_attribute": {
      "property2": "value2",
      "property3": "value3"
   },
   "key1": "value1",
   "stamp_key": "value_key"
}

Checking the snapshot files all warnings were persisted. Nevertheless on startup only one is read.

Create a new Input object that displays the number of RAID sessions by user,
the total of active sessions and also the sessions peak.

The idea is to have a quick way of knowing if we have to many RAID sessions open and also how many for a specific user.

Create a objects.disabled folder to mimic the same functionality available with inputs, validations and outputs.

Create a Handlebars helper so environment variables can be accessed for specific attribute, query, command, etc... customizations.

For better or more flexible ,index management on ElasticSearch , a feature that enables us to choose to use a daily , weekly and maybe monthly index creation.

Review all config/objects to ensure all accept a single map argument (and add documentation):

• nInput_AFPing
• nInput_CBPMSemaphores
• nInput_DB
• nInput_FMSRuleStatus
• nInput_Filesystem
• nInput_FilesystemCount
• nInput_IMMemory
• nInput_LogErrorAgg
• nInput_OSInfo
• nInput_RAIDMemory
• nInput_RunningFlows
• nInput_Semaphores
• nInput_Sessions
• nInput_Shell
• nOutput_Channels
• nOutput_EmailWarnings
• nOutput_H2
• nOutput_HTTP
• nOutput_HTTP_JSON
• nOutput_Log
• nOutput_Oracle
• nOutput_Stdout
• nValidation_AFPing
• nValidation_Semaphores
• nValidation_Yellows

Build an input to test the reach-ability of a SFTP server and check a list of folders. The attribute could have the total time of testing for compare results across time plus the result of testing each folder.

Group similar warnings displaying "plus x amount of similar warnings" configurable on the jobs arguments.

Warnings should get better housekeeping. A way to do it is to subscribe the channel with a function to clean closed warnings.

There should be a list or channel or similar with the list of active plugs (e.g. inputs, outputs. validations) for informative proposes.

Change the line if(isUnDef(meta.chSubscribe) && meta.chSubscribe == "nattrmon::warnings")
to if(isUnDef(meta.chSubscribe) || !(meta.chSubscribe == "nattrmon::warnings"))

There should be a way to get a list of all plugs (inputs, outputs and validations) and the corresponding time intervals, cron, waitForFinish, onlyOnEvent, etc...

Create a new nOuput_MVS based on openaf issue #32.

On startup of nattrmon.js all 4 main channels (Current Values, Last Values, Attributes and Warnings) don't seem to restore from the persisted corresponding snapshot files. Nevertheless their data is persisted on the snapshot files.

To allow for test/run operations on nOutput_Channels to retrieve more info from a plug execution other than just attributes the plug return should be changed to a map and dependencies changed accordingly. This would allow for the return of the warning objects created on validations, for example.

Besides the disabled folders add a new disabled/ignore option similar to git ignore.

Hello,

Having this file on .nattrmonignore :
/some/dir/config/inputs/anotherdir/a.input.pf.json
is leading to nattrmon ignoring all the inputs in the dir:
/some/dir/config/inputs/anotherdir/ instead of ignoring only a.input.pf.json in that dir.

Thanks in advance,
HD

There are new plugs for inputs, outputs and validations in YAML. A specific nInput_Init, nOutput_Init and nValidation_Init could be created in order to be possible to have init yamls which would easy configuration.

Using templify for SQLs would allow for easy building of SQL query using the previous values, last date checked, last date modified, previous value modified date, etc.

Create a generic method on the warnings object to set as notified a warning given a title.

Add the fields End Date and Duration (if no end date, should be empty) of execution fields to the existing objects for the flows executions.:
nInput_CBPMRunningFlows, nInput_RunningFlows and nInput_BPMNRunningFlows

Improve nValidation_Generic to aggregate multiple similar warnings if specified (e.g. add argument to provide a warning similarity >5%) to avoid generating hundreds of warnings if they are very similar.

Plugs actually can be described in .YAML or they can be added programmatically with .JS. Some users prefer json to yaml when writing each plug description. So it would be nice on the .YAML plug reader to also be able to read *.json files.

It's sometime hard to understand if there was a warning creation or closing. A global logging of these events would be helpful.

The nOutput_EmailWarnings could be improved to include a indication for templates to distinguish between new warnings and already notified warnings.

An improvement to nInput_Shell would be to allow to use a template for cmd (in cases the command arguments are based on other current attribute values)

It should be possible to force the execution of a plug remotely for debug proposes.

Document changes introduce by 13fe829

Create new Input to display all schedulers.

Arguments:

• status_include: receives an array of status to include (default: includes everything)
• status_exclude: receives an array of status to include (default: excludes nothing)
• name_exclude: receives an array of scheduler names to exclude (default: excludes nothing)

Columns to display:

• Category

• Name

• Status

• Next Execution Date

• Last Execution Date

• How Long Ago