Open Gupta-Pankhuri opened 3 weeks ago
Hi @Gupta-Pankhuri !
let start with the easy one ;)
The queries
is not the ACI Class but this paramenter needs to match the queries name defined in the aci-exporter config file. for example if you are probing an APIC you would say queries=controller_topsystem
and for a switch queries=node_topsystem
you can see this in this the config file: node.yaml
node_interface_infos: What is 172.16.121.2? node_interface_infos Needs to be directed to the IP address of a switch no to the APIC. I get a similar output in ,y lab if I use the APIC IP as target. The whole idea of the stack is to distribute the queries directly to the switches not to overburden the APIC.
If you deploy the stack this is taken care automatically by the prometheus config. Are you deploy the stack or just aci-exporter?
172.16.121.2 is my APIC-controllers IP, if not APIC IP address then what to use which config file to use and how to use the exporter, I'm currently running the exporter with ./build/aci-exporter -config fab1.yaml and fab1.yaml is my config file where I'm giving APIC's information.
The node
parameter must point to the switch inband or out-of-band address.
If you deploy the aci-monitoring-stack then prometheus, via service discovery takes care of this for you. Clearly ACI exporter needs to have access to the switches IP.
Also if you are planning to use the aci-exporter stand alone (or just want to understand better how it works) you should probably take a look at its repo https://github.com/opsdis/aci-exporter
I was refering to this repo only and was using example-config.yaml file but it was giving same data and service discovery is already happening that data is available as I attached the screenshots, I just want to run exporter from my end where prometheus is deployed and monitor APIC through it...If you're comfortable can we connect?
Broadly speaking if you want to re-use the queries from this repo if a query start with
node_xxxxxx
then it has to be sent to the node ip. i.e. probe?target=fab2&node=192.168.68.61&queries=node_interface_info
where 192.168.68.61
is the management IP of the switch. node
then is an APIC query and you do not need to add node
in the probe URL.If you want to use just the exporter as said please read the documentation of the aci-exporter and start getting familiar with it with simple queries.
You could try to follow this example as well: https://github.com/datacenter/aci-monitoring-stack/blob/main/docs/development.md#aci-exporter-and-prometheus
Where I try to guide you through configuring the exporter.
The aci-exporter repo also has some examples that you could use get yourself familiar with.
According to the mentioned example I just didn't understand that where exporter needs to be run and setup should be done because prometheus and ACI node are different in my case so I was running exporter and config files from prometheus node. Therefore, curl "http://aci-exporter-ip:9643/probe?target=fab1&node=192.168.68.8&queries=node_interface_info" in this API I'm confused if in aci-exporter-ip I should enter prometheus IP or APIC's IP if I'm running exporter on prometheus node with config file consisting of APIC's detail.
If you look at this section in the read.me https://github.com/datacenter/aci-monitoring-stack?tab=readme-ov-file#aci-object-scraping you can see that prometheus polls ACI exporter so in the example:
"http://aci-exporter-ip:9643/probe?target=fab1&node=192.168.68.8&queries=node_interface_info"
prometheus is the one what would make this HTTP request.
aci-exporter-ip
is the IP address of the server where the aci-exporter
is running target
is the fabric name as defined in theaci-exporter
config node
is the IP address of the switch you are polling the data from. prometheus
can run on the same server as the aci-exporter-ip
or anywhere really
Hi,
I am currently deploying a solution to fetch ACI metrics and have successfully retrieved data for the sd(service discovery). However, I am encountering issues when attempting to fetch data for the following classes:
--> node_interface_info: It seems that I am receiving duplicate data. --> topSystem: I am unable to retrieve any metrics for ACI. --> SD - data is available
Could you please provide guidance on where I might be making a mistake and suggest potential solutions for these issues?