* fix: Check HC counters individually
Vendors are very inconsistent returning ifXTable values.
Some only implement some of them (seemingly randomly).
Others return them, but they are always 0.
This attempts to fix that by checking each value individually.
* A couple of small cleanups
* Change the logic, The (if < hc) logic will be messed up when the counters roll.
Just make sure it is not 0 or an empty string.
* Initial support for graphite
* Enable graphite include
* fixed typos.
* Fixed port naming to graphite metric names
* Added documentation for Graphite
* added documentation for graphite
* fixed style
* replace / with _ in interface names for graphite
* sets default graphite port
* adds rrd-name tags to the metric, otherwise metrics get lost
* add suggested storage schema for graphite
* add whitespace
* Updated to show some output
* bad english
* Use HC counters even if one direction is reporting as 0
* Change to >=
* Change this to a for loop
Treats each item individually. I left it as >, because if the HC value is 0, then the normal value is 0. And it no longer affects other values as we treat each one individually.
* Change check again...
* Update test for HC being used
* Last time?
* Typo
* feature: Allow customisation of rrd step/heartbeat when creating new rrd files
* revert defaults
* added docs + webui config option
* Move RrdDefinition to an Object to make them easier to create and remove the possibility of typos.
* Fix style/lint issues and missing use statements
* 3 more missing use statements
* updated doc + moved schema file
* feature: Added ability to switch port polling mode
* update, speed things up slightly by walking certain tables
* more fixes
* updates
* updates
* update to exclude ports that are down
* wrong variable used
* update to set previous port status when null
* small update to port disco
* New event log severity model
* Revert "New event log severity model"
This reverts commit 0783527af8.
* New event log severity model
* Fixed spaces vs tabs
* Learning the format requirements
* Learning the format requirements
* Moved the colours in to the stylesheet
* change SQL file to stop conflicting
* refactor log_event so severity isn't last, but reference is.
* keep having to move the file due to new PR conflicts
* Revert "keep having to move the file due to new PR conflicts"
This reverts commit f3030e7218.
* extended support to dashboard and eventlog page
* formatting
Use data_update instead of rrd_update/rrd_create and influx_update
Centralize rrd file check so we can check against a remote rrdcached server too
Use rrd_name() to generate rrd file names
The previous ifIndex based port mapping schema did detect deleted ports
by checking ports in the DB against the ports currently polled via SNMP
which breaks when using another port mapping schema. This commit fixes
problem by checking known and found ports by their port_id.
Signed-off-by: Maximilian Wilhelm <max@rfc2324.org>
