if you want, we can have a short call together to review this requirement. I guess that with 15-20 min we should have enough time to create the alert that you need. Please send me an email at raul.gonzalez@ibm.com to arrange the session.
Original Message:
Sent: Fri September 15, 2023 06:49 AM
From: AKHIL Raj
Subject: Interface CRC error monitoring using Sevone
Thank you @Raul Gonzalez for this solution . We are trying to bring up the Sevone thresholds similar to the existing configuration ( trigger alert when condition met for 30 min out of 1hr ) . I understood we have to use total aggregator instead of Time over threshold which means we can't give a condition with sliding window. I will configure the same in policy and test it.
And also Is it feasible in sevone to create a total - aggregation with Sliding window ( trigger alert when condition met for xx min out of yy minute )feature ?. If it can be doable I will open a support request for this.
------------------------------
AKHIL Raj
Original Message:
Sent: Fri September 15, 2023 06:30 AM
From: Raul Gonzalez
Subject: Interface CRC error monitoring using Sevone
Hi Akhil
unfortunately that will not work because the default behaviour on COUNTER OIDs is to use the number/sec. Unless you use a smaller threshold such as 500 CRC errors / 300 seconds (if you monitor every 5 minutes) = 1.66 as threshold then that approach would work.
Normally for COUNTER metrics where you are interested not in the rate but in the absolute number, we would recommend using TOTAL as aggregation.
------------------------------
Raul Gonzalez
Software Networking Solutions Architect
IBM
Brighton, UK
Original Message:
Sent: Fri September 15, 2023 06:23 AM
From: AKHIL Raj
Subject: Interface CRC error monitoring using Sevone
Thank you Raul Gonzalez, for your quick assistance .
I understood the point , I have configured the threshold condition using aggregation time over threshold ( if CRC 500 number > 30 min out of 1hr ). Does that aggregation also behave like average ?
------------------------------
AKHIL Raj
Original Message:
Sent: Fri September 15, 2023 06:06 AM
From: Raul Gonzalez
Subject: Interface CRC error monitoring using Sevone
Thank you Akhil
I think I know where the issue is, when working with COUNTERS, SevOne shows you the unit/second, not the unit by itself, that's why on the top screenshot you can see that there has been 500-600 errors, whereas in SevOne you will see 1 error/second.
If you want to see the absolute number of errors you have to change the aggregation from 'average' (or nothing) to 'total', that should allow you to see very similar values. This also affects the way we create alerts, you should select aggregation 'total' rather than 'average'.
Example with traffic, top graph is using bytes/sec, second graph is using total bytes
On alerting policies we should select Total rather than Average:
------------------------------
Raul Gonzalez
Software Networking Solutions Architect
IBM
Brighton, UK
Original Message:
Sent: Fri September 15, 2023 05:56 AM
From: AKHIL Raj
Subject: Interface CRC error monitoring using Sevone
Report from another NMS - its in UTC time zone y axis in k(thousands).
------------------------------
AKHIL Raj
Original Message:
Sent: Fri September 15, 2023 04:35 AM
From: Raul Gonzalez
Subject: Interface CRC error monitoring using Sevone
Hi Akhil
short answer: yes, it is possible.
long answer: one of the problems we have with vendors is that they don't normally share CRC errors using standardised protocols such as SNMP. Some times they try (like Cisco did with the OID .1.3.6.1.4.1.9.9.45.1.1.1.20) but most of them fail miserably :). So what options do we have? as you might possibly know, SevOne is able to collect any time series data, as long as that data is available in some kind of format. I understand that you are using some kind of CLI command to get the CRC errors, well, with SevOne you can do the same, execute that command (either using our nocode platform, or if you prefer using a custom script) and then ingest that data back into SevOne.
Similar principles can be found here: https://community.ibm.com/community/user/aiops/viewdocument/monitor-ibm-cloud-metrics-using-the?CommunityKey=fe9d91df-352c-4846-9060-189fd98d00ca&tab=librarydocuments where we are monitoring data coming from APIs.
------------------------------
Raul Gonzalez
Software Networking Solutions Architect
IBM
Brighton, UK
Original Message:
Sent: Thu September 14, 2023 12:55 PM
From: AKHIL Raj
Subject: Interface CRC error monitoring using Sevone
Can someone put some light on this subject. We have a requirement to monitor the CRC interface errors . Currently we tried to create it using fcs error counter but the value showing up in sevone is very less while averaging compare to the existing monitoring tool.
We are looking for some help to do a effective crc monitoring,
------------------------------
AKHIL Raj
------------------------------