* feature: Add a 'session' sensor type
This allows things like the SRX5800 SPU's to be graphed as they expose current and maximum usage in a table.
The intent is to contain things like TTY sessions, flow counters from other vendors, or other resources with support a finite number of simultaneous users or connections.
* Include the node description in the descr field
Chassis clusters use the same FPC's
* Track PanOS session count
* Add JunOS CP Flows
* Missed file containing overview
* Missed a spot
* remove custom sessions sensor and use count
* Update junos.yaml
* Update panos.yaml
* Dynamic entPhysicalIndex values for YamlDiscovery
* Use degree symbol when displaying temperature values
* Adjust SNMP options for new snmp_get_multi()
The OS polling module for mrv-od was using snmpget commands with
-Ovqs, rather than -OQUs, which meant that the equal sign was
stripped in the snmpget output. But that equal sign is now used
as the explode() delimiter so it broke the construction of the
returned array for the snmp_get_multi() function.
Additionally, I removed the NBS-CMMC-MIB:: prefix which is no
longer returned by snmp_get_multi() as part of the OID. Now,
the mrv-od.inc.php OS polling module is retrieving data from
positions in the array with actual content.
* Convert mrv-od to yaml, improve sensors discovery
Converted sensors discovery for mrv-od to yaml, retaining all
functionality for pre-cache, dbm, and temperature. Replaced specific
SFF-8472 Digital Diagnostics state sensors with generic ones, given that
we are now reading out detailed values through new dbm, current, voltage
sensors.
Reasoning: the generic Digital Diagnostics state sensors report state
diagsOk, or else the worst severity level measured among all of the
Digital Diagnostics values for that optical data point. You'd really
only want to set a generic "alarm" so the box can report that it
thinks something is wrong, whereafter a human can look at the graphs
and/or the CLI to draw their conclusions about the alarm. Otherwise,
things would quickly become too spammy as you are essentially duplicating
every sensor across both regular measurements and state measurements.
Further changes:
- Increase maximum number of detected PSUs from 4 to 8 per the MIB.
- State sensor to detect (in)sufficient power budget for all modules.
- Use skip_values to hide state sensors with the notSupported value.
- Use skip_values to hide sensors which are unsupported per the MIB,
signified by the device returning special values -1 or -2147483648.
- Add dBm, temperature "over" graphs to the mrv-od device header bar.
* Add mrv-od entity-physical support
MRV does not use the ENTITY-MIB data format, so we had to reconstruct
the inventory from their proprietary format using the following tables
from their NBS-CMMC-MIB:
- nbsCmmcChassisTable
- nbsCmmcSlotTable
- nbsCmmcPortTable
Additionally, since sensors didn't have their own ifindex in the
NBS-CMMC-MIB, I went ahead and appended the last OID digit to
the "containing" ifindex create a unique number. The same was
done for the sensors in the yaml-discovery.
* Test data for mrv-od, fix 2* pre-commit.php issues
* Fix entPhysicalIndex returning 0 instead of null
Some unit test breakage was caused by 62167ebcb which changed
the code to return 0 instead of null when {{ $index }} was
set as the entPhysicalIndex in the YamlDiscovery. So now we
detect if it returns some value which may be interpreted as
false, and if so we set null rather than 0.
* Added configuration options to aggregate input and output bits before making 95th percentile billing calculations
* Changed aggregate to per-bill instead of global. Added config options for making aggregate the default selected option. Refactored out mres() calls in touched files. Changed to Config::get where appropriate.
* Fixed documentation typo
* Fixed scope of aggregate default config option to be under billing
* Fixed getRates issue in Quick Graphs section of 95th percentile bill
* Added configuration options to aggregate input and output bits before making 95th percentile billing calculations
* Changed aggregate to per-bill instead of global. Added config options for making aggregate the default selected option. Refactored out mres() calls in touched files. Changed to Config::get where appropriate.
* Fixed documentation typo
* Fixed scope of aggregate default config option to be under billing
* Fixed getRates issue in Quick Graphs section of 95th percentile bill
* Fixed formatting
* Fix for RouterOS LLDP discovery
Only the first nine neighbors were correct on RouterOS - missing hexdec() led to the remaining neighbors being linked to the wrong ports. Ex. port 10 in hex was incorrectly linked to port 10 decimal when it should be 16 in decimal
* Added test data for MikroTik RouterOS CRS317 switch
* Added configuration options to aggregate input and output bits before making 95th percentile billing calculations
* Changed aggregate to per-bill instead of global. Added config options for making aggregate the default selected option. Refactored out mres() calls in touched files. Changed to Config::get where appropriate.
* Fixed documentation typo
* Fixed scope of aggregate default config option to be under billing
* Instructions for logstash
How to configure logstash to send syslog to LibreNMS
* remove silly ssh syntax tags
* Fixed date formatting issue for afternoon times
service_alerting_enable no will properly disable alerting (may be set globally or per node)
service_alerting_frequency will no properly control frequency (set globally the same, via db is best)
* fix client count polling, add ap count polling.
* Set low warning/critical thresholds for AP count sensors
* Add Aruba AP Count Warning and Critical Alerts to collection, lint alert_rules.json
* add newline at end of alert_rules.json
* add/update tests/data/arubaos*.json test data
* Support for system APP_LOCALE
* Start preferences re-write
* port 2fa form
* Working user preferences
* Language user preference
* Don't look up locale from the DB every request
* Device list working
* Deny demo user middleware
* Finish password changing
* remove used resource methods
* remove leftover use
* warn that translation is incomplete
* fix style
RunFollowUp checks if there is a custom sql query attached to
the alert and uses this before invoking GenSQL to build the query.
However, the custom query column was not included when selecting
alerts so the GenSQL function was always invoked.
This patch includes the alert_rules.query in the loadAlerts function
used by RunFollowUp.
* Fix Cisco MAC acounting discovery, polling and HTML templates.
* Remove empty code.
* Provide test data for cisco-mac-accounting module.
* Added module test dependency.
* Refactor LibreNMS service
add ping
* services ported
remote legacy stats collection
* alerting
* implement unique queues
* update discovery queue manager
* remove message
* more cleanup
* Don't shuffle queue
* clean up imports
* don't try to discover ping only devices
* Fix for discovery not running timer
* Update docs a bit and and add some additional config options.
Intentionally undocumented.
* Wait until the device is marked up by the poller before discovering
* Handle loosing connection to db gracefully
* Attempt to release master after 5 db failures
* Sleep to give other nodes a chance to acquire
* Update docs and rename the doc to Dispatcher Service to more accurately reflect its function.
* add local notification
* Collect Appliance serial number from Sophos device
Add serial number read. Added on polling session instead of discovery because the S/N changes in case HA functionality is enabled, and there is a switch between active and standby.
* Update cyberoam-utm.inc.php
* Fix Siklu Version/Serial
Siklu allows for two software images. We must first determine which SW bank is "running" in order to report the correct version.
Also adding correct Serial Number detection for Siklu devices.
* Update siklu.inc.php
* Update siklu.snmprec
* Create siklu.json
