Tidying that up:
apCurrentNoise -92
apSysStatBssid{apSysStatBssid="07:8B:67:FF:4F:35"} 1
and those will have job/instance labels added by Prometheus.
The main problem will be that ths bssid timeseries come and go when the
bssid changes - the timeseries with the old bssid will be seen for 5
minutes (the default lookback delta). So it's not a case so much of
checking the AP is on the *same* bssid as it was before, but that it's
*not* been seen on a different bssid.
>From the top of my head, and completely untested, here is a starting point:
abs(apCurrentNoise - apCurrentNoise offset 1m) >= 5 unless on (instance)
count by (instance) (apSysStatBssid) > 1
This should suppress the alert if the AP has changed bssid within the last
5 minutes.
abs(pCurrentNoise - apCurrentNoise offset 2m) >= 5
On Tuesday, 11 March 2025 at 19:26:20 UTC Matthew Koch wrote:
> So The BSSID is the constant and I want to know when the apCurrentNoise
> changes by -/+5 but only if I am connected to the same access point
> (BSSID). If the BSSID changes and the Noise level changes then its not as
> big of a deal. It's only if I am connected to the same BSSID and in a new
> polling interval the noise level changes. # HELP apCurrentNoise Background
> noise level (dBm) of the 802.11 device. -
> 1.3.6.1.4.1.11.2.14.11.6.4.5.1.1.21.6 # TYPE apCurrentNoise gauge
> apCurrentNoise -92 # HELP apSysStatBssid MAC address of the AP that the
> HPE501 is Associated to. - 1.3.6.1.4.1.11.2.14.11.6.4.5.1.1.1.1.31.3 # TYPE
> apSysStatBssid gauge apSysStatBssid{apSysStatBssid="07:8B:67:FF:4F:35"} 1
>
> On Thursday, March 6, 2025 at 1:35:22 PM UTC-5 Brian Candler wrote:
>
>> If you can show complete examples of the metrics (the noise level one and
>> the one you want to correlate to, including all the labels) then it may be
>> possible to show you how to join them.
>>
>> On Thursday, 6 March 2025 at 15:43:25 UTC Matthew Koch wrote:
>>
>>> This is used to track how a wireless client is connected to a specific
>>> access point. What I am trying to determine is when a client has endured a
>>> significant amount of interference so when I read the noise level from the
>>> client I want to report when that metric has changed significant but I only
>>> want to do this if the mac address which is collected at the same time is
>>> unchanged. We need multiple readings to determine this so when I compare
>>> one point in the time series db to another I use the MAC address for
>>> comparison. the challenge is I cannot do this with the label.
>>>
>>> On Wednesday, March 5, 2025 at 3:39:35 PM UTC-5 Brian Candler wrote:
>>>
>>>> Sorry, I don't understand - what sort of comparison do you want to do
>>>> on a MAC address?
>>>>
>>>> On Wednesday, 5 March 2025 at 19:23:09 UTC Matthew Koch wrote:
>>>>
>>>>> In Grafana I am trying to use the metric to do a comparison, but I
>>>>> cannot do it because it is a label and not a value. SNMP returns a value
>>>>> of
>>>>> 1 with the label
>>>>>
>>>>> On Wednesday, March 5, 2025 at 12:56:42 PM UTC-5 Brian Candler wrote:
>>>>>
>>>>>> OK. I checked this. regex_extracts is only from when you want to
>>>>>> extract a metric *value* (i.e. a floating-point number) from text, for
>>>>>> some
>>>>>> dodgy MIBs which respond with things like temperature as a string value.
>>>>>>
>>>>>> Since you already have apSysStatBssid containing the MAC address as a
>>>>>> string label, what exactly do you want to do with it?
>>>>>>
>>>>>> On Wednesday, 5 March 2025 at 16:43:08 UTC Matthew Koch wrote:
>>>>>>
>>>>>>> I've tried the variety of Regex and still it doesn't return any
>>>>>>> data. It's interesting because nothing comes back in the output and it
>>>>>>> doesn't error as well.
>>>>>>>
>>>>>>> On Wednesday, March 5, 2025 at 3:57:22 AM UTC-5 Brian Candler wrote:
>>>>>>>
>>>>>>>> Start with regex: '(.*)'
>>>>>>>>
>>>>>>>> If that works, then you know it's just the regex at fault. I'd
>>>>>>>> start by trying to double-backslash, i.e.
>>>>>>>>
>>>>>>>> - regex: '([\\w]{2}:[\\w]{2}:[\\w]{2}:[\\w]{2}:[\\w]{2}:[\\w]{2})'
>>>>>>>>
>>>>>>>> The spec for Go's regex language (RE2) is here:
>>>>>>>> https://github.com/google/re2/wiki/syntax
>>>>>>>>
>>>>>>>> I'm not sure if a backslash character class is allowed inside a
>>>>>>>> square-bracket character class. So you could try:
>>>>>>>>
>>>>>>>> - regex: '((\\w){2}:(\\w){2}:(\\w){2}:(\\w){2}:(\\w){2}:(\\w){2})'
>>>>>>>> - regex:
>>>>>>>> '([[:alnum:]]{2}:[[:alnum:]]{2}:[[:alnum:]]{2}:[[:alnum:]]{2}:[[:alnum:]]{2}:[[:alnum:]]{2})'
>>>>>>>>
>>>>>>>> or something simpler like:
>>>>>>>>
>>>>>>>> - regex: '([a-fA-F0-9:]{17})'
>>>>>>>>
>>>>>>>> On Tuesday, 4 March 2025 at 22:11:04 UTC Matthew Koch wrote:
>>>>>>>>
>>>>>>>>> I'm trying to extract the MAC address out of what is returned by
>>>>>>>>> SNMP exporter. The Regex I am using seems to work in general when I
>>>>>>>>> used an
>>>>>>>>> online Regex tester but for some reason it's not working in SNMP
>>>>>>>>> exporter.
>>>>>>>>>
>>>>>>>>> SNMPExporter config:
>>>>>>>>>
>>>>>>>>> - name: apSysStatBssid
>>>>>>>>> oid: 1.3.6.1.4.1.11.2.14.11.6.4.5.1.1.1.1.31.3
>>>>>>>>> type: PhysAddress48
>>>>>>>>> help: MAC address of the AP that the HPE501 is Associated
>>>>>>>>> to. - 1.3.6.1.4.1.11.2.14.11.6.4.5.1.1.1.1.31.3
>>>>>>>>> regex_extracts:
>>>>>>>>> Test:
>>>>>>>>> - regex:
>>>>>>>>> '([\w]{2}:[\w]{2}:[\w]{2}:[\w]{2}:[\w]{2}:[\w]{2})'
>>>>>>>>> value: '$1'
>>>>>>>>>
>>>>>>>>> This is what I typically get data wise without the regex config:
>>>>>>>>>
>>>>>>>>> # HELP apSysStatBssid BSSID of currently connected accesspoint -
>>>>>>>>> 1.3.6.1.4.1.29456.3.2
>>>>>>>>> # TYPE apSysStatBssid gauge
>>>>>>>>> apSysStatBssid{apSysStatBssid="6A:56:E3:7A:85:47"} 1
>>>>>>>>
>>>>>>>>
--
You received this message because you are subscribed to the Google Groups
"Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion visit
https://groups.google.com/d/msgid/prometheus-users/d82efd86-0f8e-4405-9c2d-be2925606da5n%40googlegroups.com.
