[prometheus-users] Re: snmp_exporter related configurations.

2023-02-22 Thread Brian Candler
On Wednesday, 22 February 2023 at 12:07:31 UTC+8 chintaman...@gmail.com 
wrote:
1.  snmp_exporter should be installed on prometheus server/VM or need 
dedicated server/VM ?

Either works fine, so I suggest you start simple and only make it more 
complex if you ever find you need to.

I've always installed snmp_exporter on the same host as prometheus itself, 
since both prometheus and snmp_exporter use relatively little resources; 
but it depends on the amount of scraping you're doing.

One reason to install snmp_exporter on a separate host would be for network 
reachability reasons: that is, where the remote targets are not directly 
reachable from the prometheus server, but you can install snmp_exporter in 
a location where prometheus can talk to snmp_exporter, and snmp_exporter 
can talk to the target hosts.

(In that sort of situation, another option is to have a second instance of 
prometheus in 'agent' mode, colocated with snmp_exporter; the prometheus 
agent then pushes metrics to a central prometheus server using remote write)
 
2. From how many nodes , snmp_exporter can collect metrics ?

As many as you like, within the resources you have.
 
3. In case of overload, how can multiple snmp_exporter can connect with 
single (or multiple) prometheus servers.?

It's very flexible, you just configure it how you like.

If you need multiple prometheus servers each scraping a subset of targets 
(sometimes called "sharding") then you could have those prometheus servers 
all scraping the same snmp_exporter. However, since snmp_exporter is so 
small and cheap, I would be inclined to give each prometheus server its own 
snmp_exporter instance.

The other way round, you can have a single prometheus server talking to 
multiple snmp_exporters, by having multiple scrape jobs - but since 
snmp_exporter can already make use of multiple cores, this is unlikely to 
have much benefit unless those snmp_exporters are located on different 
hosts as mentioned above..
 
4. where exactly is auth: for targets? (in snmp.yml, or other file)

In snmp.yml
 
5. Is it essential to build using generate?

No, but it can be a bit fiddly to hack snmp.yml.  You might find it's 
helpful to generate snmp.yml out of parts, and cat them together to form 
the final file to be read by snmp_exporter.

A simple trick to take the existing snmp.yml and add authentication is here:
https://groups.google.com/g/prometheus-users/c/WJAJ07fIlDw/m/RlnW3P3UBAAJ
 
I saw few references, downloaded snmp_exporter (confirmed that prometheus 
server is getting SNMP response from targets), did initial configurations, 
added community string in auth: of snmp.yml , added config to 
prometheus.yml  
Still unable to reach to targets. SO wan to recheck steps by referring any 
detailed document for snmp_exporter

I suggest you test snmp_exporter directly using curl, which keeps 
prometheus out of the question until you have SNMP communication working 
successfully.  I also suggest you start with if_mib which is the simple 
basic one that pretty much everything supports.

Maybe this example will help:
https://nsrc.org/workshops/2022/rwnog/nmm/netmgmt/en/prometheus/ex-snmp-exporter.html

-- 
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 prometheus-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/prometheus-users/fc76127d-16aa-4b3b-9436-580513342789n%40googlegroups.com.


[prometheus-users] Re: snmp_exporter related configurations.

2023-02-21 Thread chintaman...@gmail.com
Error details, as observed in UI.
An error has occurred while serving metrics: error collecting metric 
Desc{fqName: "snmp_error", help: "Error scraping target", constLabels: {}, 
variableLabels: []}: 
error getting target 192.168.1.2: request timeout (after 3 retries)

On Wednesday, February 22, 2023 at 9:37:31 AM UTC+5:30 
chintaman...@gmail.com wrote:

> Hi All,
> Are there any in-depth detailed documents and videos, which describes 
> snmp_exporter configurations?  including some architecture details, like, 
> 1.  snmp_exporter should be installed on prometheus server/VM or need 
> dedicated server/VM ?
> 2. From how many nodes , snmp_exporter can collect metrics ?
> 3. In case of overload, how can multiple snmp_exporter can connect with 
> single (or multiple) prometheus servers.?
> 4. where exactly is auth: for targets? (in snmp.yml, or other file)
> 5. Is it essential to build using generate?
>
> I saw few references, downloaded snmp_exporter (confirmed that prometheus 
> server is getting SNMP response from targets), did initial configurations, 
> added community string in auth: of snmp.yml , added config to 
> prometheus.yml  
> Still unable to reach to targets. SO wan to recheck steps by referring any 
> detailed document for snmp_exporter
>
> Regards,
>  
>

-- 
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 prometheus-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/prometheus-users/2e71909b--4918-bb9a-74d503265ebfn%40googlegroups.com.