The NVIDIA application obtains its metrics from nvidia-smi dmon. Its built-in help (nvidia-smi dmon --help) says that the pviol and tviol columns stand for "Power and Thermal Violations". Therefore the metric should not be called Thermal Violation Percentage, but Power Violation Percentage.
* Clean broken VRF lite code
* Change DB table for route discovery
* Add VRF simple support
* add port_id to db and discovery
* static-fy the translation arrays
* sort and search cleaning
* Sorting refactor and validation
* formatItem shortened
* Handle ifIndex==0 meaning no next hop defined (MPLS)
* Sync all create/updates
* purge in daily
* remove old route table
* get rid of inetCidrRouteNextHop_device_id
* fix wonky column orders
* add route snmprec
* fix sorting by interface
* Move to new config
* rename to route the new table
* Properly display ipv6 compressed addresses
* Translation before merge ./lnms translation:generate
* Update manifest
* Device Cache
also some additional device related models
* attribs from cache
* replace common attribute functions
* remove legacy cache usage
tidy up some collection manipulation
remove some unused or single use functions
* cleanup some items
* always return a device, to prevent bugs
* clear device cache when testing after each test
* fix double assignment
* Clean up function to take advantage of null object
* - Adjust Alerting Query Builder logic to not use boolean/radio buttons for 'packet_loss_' macros
- Add nagios check script for "check_oracle", which fails on the fallback logic because it does not accept IP/hostname via the '-H' flag
* Update check_oracle.inc.php
* allow user specific themes
* add missing newline
* use global function getDefinitions
* some fixes
* .
* travis fix
* .
* Optimize preference loading
Aka, don't run multiple sql queries per page load, now it is 0-1 queries (for preferences)
* Add a default option for user preferences
* Remove unused code
* more
Only call snmpwalk once, instead of 20, to avoid havin to face up to 20 times the timeout. Decreased drastically polling time on some Cisco 887 / 897 devices
* Corrected filtering of devices under maintenance
* Clarify wording of disable/alert device settings and dependency between the two settings
* Disable "Ignore alerts" checkbox according to initial setting of "Disable polling"
* Correct initial enabling/disabling of 'Ignore alerts'
* Remove not functional js code
In the 'slas' table, the column 'opstatus' is of type int. The current
comparison does not work correctly - in the web-interface all sla-probes are
displayed as 'down', even if the status = 0 in the table (0 is OK).
* fix restricted application access for normal user
* travis fix
* use of Permissions class
* travis fix
* Update default.inc.php
* Update overview.inc.php
* travis fix
* - update yzyxel mibs
- update definitions für ies series
- slot state and alarm support via yaml
* more sensors
* add new device Zyxel IES4206
* restore original ies files
* revert function stuff
* update to latest official mib
* rebase
additional translatable strings: Filter Settings in settings page and Global Search on all pages.
remove duplicate fr key
also, language file signatures were wrong as I had a testing version of the vue generator
* Unbundle javascript language files
automatically generated for each install
* update docs
* update laravel-vue-i18n-generator
a newer (unreleased) verison is need for English fallback to work
* user deactivation feature
* update db_schema.yaml
* travis fix
* readd sqlfile with alter statement
* ..
* revert force push
* combine all queries
* fix query
* user enable/disable only visible on mysql authorization
* Update form.blade.php
* Update index.blade.php
* disable 'enabled' on own profile
* bootstraping checkboxes
