If you need to post any text longer than a few lines, please use a pastebin service such as https://p.libren.ms using non-expiring pastes.
I’m trying to monitor a bunch of Arista infrastructure but am getting a tonne of alerts for RX/TX power and RX/TX bias from each switch. It’s hard to know if these alerts are valid or not - looking at Update LibreNMS to grab Arista optic high/low sensor values · Issue #7078 · librenms/librenms · GitHub it seems as though Libre has support for pulling threshold values from Arista switches. However, when I compare threshold values for two identical switches I can see they’re different, which would suggest they have been calculated on the fly when the device was discovered?
[OK] Composer Version: 2.4.4
[OK] Dependencies up-to-date.
[OK] Database connection successful
[OK] Database Schema is current
[OK] SQL Server meets minimum requirements
[OK] lower_case_table_names is enabled
[OK] MySQL engine is optimal
[OK] Database and column collations are correct
[OK] Database schema correct
[OK] MySQl and PHP time match
[OK] Active pollers found
[OK] Dispatcher Service is enabled
[OK] Locks are functional
[OK] No active python wrapper pollers found
[OK] Redis is unavailable
[OK] rrdtool version ok
[OK] Connected to rrdcached
The RX/TX power do seem to match the switches values upon examination.
It’s just the RX/TX Bias that seem to be a bit all over the place. Do you know where these can be examined on the switch?
The values in LibreNMS don’t match the results I’m reading out from the switch…
Looks like I’m going to have to amend them manually in device health settings, unless you know of a better way?
LON-CO3-SW-P-SFP-04-CL#show interfaces ethernet 18 transceiver dom thresholds
Ch: Channel, mA: milliamperes, dBm: decibels (milliwatts),
C: Celsius, V: Volts, NA or N/A: not applicable.
Port 18
Last update: 0:00:00 ago
High Alarm High Warn Low Warn Low Alarm
Value Threshold Threshold Threshold Threshold Unit Indicator
---------------------------------------------------------------------------
Temperature 31.66 80.00 75.00 0.00 -5.00 C
Voltage 3.35 3.60 3.50 3.10 3.00 V
Ethernet18 (Channel 1)
TX bias current 0.00 90.00 85.00 10.00 5.00 mA ALARM
Optical TX power -30.00 0.00 -1.00 -6.00 -7.00 dBm ALARM
Optical RX power -30.00 0.00 -1.00 -14.00 -15.00 dBm ALARM
The DOM TX Bias values in LibreNMS are:
High: 0.09
High warn: 0.085
Low warn: 0.01
Low: 0.005
Typing those values made me realise that Libre is doing a conversion to the values it’s reading out from the switch. Judging by the link you provided this may be being done incorrectly?
Actually my apologies; looking at those alert thresholds, the alerts are being generated correctly by LibreNMS.
My problem is these alerts are being generated for pretty much every port in every arista switch we have here which is incredibly noisy. I think I’m going to have to turn them off.
You can try to keep at least RX power at first, since an alert on this would lead to packet loss. Do you have alerts on all your ports for rx power too?