[Bug] Portal dashboard fails requesting information
Closed this issue · 7 comments
Review Portal dashboard environments requested.
pre-int... is requested instead of the environment in use.
I leave here a log from @sinaure
November 21st 2019, 16:37:39.026 tng-sec-gtw 192.168.2.36 - - [21/Nov/2019:15:37:39 +0000] "GET /api/v3/monitoring/graphs/d-solo/sp_dash/sp?orgId=1&panelId=3&var-id=4fd837f6-5f49-11dc-a9ed-000ea629289b&var-entity=vm&var-env=pre-int-vnv-bcn.5gtango.eu%3A4fd837f6-5f49-11dc-a9ed-000ea629289b&theme=light&refresh=10s HTTP/1.1" 200 23950 "http://192.168.1.181/dashboard" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/78.0.3904.97 Safari/537.36" "-"
Hi, as commented yesterday:
- Portal takes the graph name with all its attributes. Portal is not able to predict that var-env will be
pre-int-vnv-bcn.5gtango.eu%3A4fd837f6-5f49-11dc-a9ed-000ea629289bwith all that hash. The maximum I can do is to change thepre-int-vnv-bcn.5gtango.eupart for the location of the Portal but this will not solve anything with that following hash. - Regarding the
d-solo/sp_dash/sp?part which is also static, that is how the graphs of all Grafanas I've seen (SP and V&V ones) define the URL to get the graphs. So I see nothing to change in there for Portal while this is like this.
Examples of graphs for VnV Host CPU usage from Grafana in SP and in V&V:
SP: https://pre-int-sp-ath.5gtango.eu/api/v3/monitoring/graphs/d/sp_dash/sp?orgId=1&from=1574738703848&to=1574760303848&var-id=341f4d56-8e66-cfae-6fb3-1143fff11091&var-entity=vm&var-env=pre-int-sp%3A341f4d56-8e66-cfae-6fb3-1143fff11091&panelId=17&fullscreen
@pkarkazis please can you take a look to my last comment?
From an exchange with panos
Panos Karkazis @pkarkazis 7:54
Hello @sinaure,
I saw your issue in son-monitor (sonata-nfv/son-monitor#138) but I am not sure that I understood your point. Could you please provide me a more detailed description of the problem? For example, where the problem occurred, in which env did you notice it etc ?
aureliano sinatra @sinaure 09:35
Hi Panos, I've open this issue in duplicate with tng-api-gateway I am not sure is related with son monitor... Anyway this is the scenario:
I've installed the VnV on a VM in our 'EasyglobalMarket' dev environment (out of 5Gtango network)
we are of course in vpn
test plan execution stuck between tng-api-gateway and tng-vnv-planner
on the log stack I see
a reference to pre-int-vnv-bcn.5gtango.eu
my question is if I am using a local environment for the VnV and a remote SP (Aveiro in this case)
why do I have pre-int-vnv-bcn.5gtango.eu in the callback?
looks something hardcoded
I should have instead my internal IP
Panos Karkazis @pkarkazis 13:26
This query is originated from the portal asking a graph from monitoring (via tng-api-gtw), I suppose that the portal asks monitoring charts regarding the machine that hosts the vnv. I think that the most appropriate person to ask is @anapolg, probably you must use an env variable...
@pkarkazis this is the request for both SP
${ this.config.baseSP }${ this.config.graphs }/d-solo/sp_dash/sp?orgId=1&` +
`panelId=${ panelId }&` +
`var-id=341f4d56-8e66-cfae-6fb3-1143fff11091&` +
`var-entity=vm&` +
`var-env=pre-int-sp%3A341f4d56-8e66-cfae-6fb3-1143fff11091&` +
`theme=light&` +
`refresh=${ this.refreshRateGraphs }`
var-env= ${ window.location.origin }%3A${variable}& +
//pre-int-vnv-bcn.5gtango.eu%3A4fd837f6-5f49-11dc-a9ed-000ea629289b&
Request: curl -X GET "http://pre-int-vnv-bcn.5gtango.eu:8000/api/v2/prometheus/metrics/name/vm_mem_perc"
Response:
{
"metrics": {
"resultType": "vector",
"result": [
{
"metric": {
"__name__": "vm_mem_perc",
"exported_instance": "pre-int-vnv-bcn.5gtango.eu:4fd837f6-5f49-11dc-a9ed-000ea629289b",
"exported_job": "vm",
"instance": "pushgateway:9091",
"job": "pushgateway",
"resource_id": "4fd837f6-5f49-11dc-a9ed-000ea629289b"
},
"value": [
1574940909.247,
"50.31"
]
}
]
}
}
Filter by name and the first part of exported_instance