Splunk observability scaler #6534
Conversation
| @@ -0,0 +1,207 @@ | |||
| //go:build e2e | |||
There was a problem hiding this comment.
This looks like a cloud service so we will need to create an account. Is Splunk Free enough to exceute this e2e test?
There was a problem hiding this comment.
Unfortunately, no. The Splunk Observability Cloud (former SignalFX) is a different product.
There was a problem hiding this comment.
We would need an account then, @sschimper-splunk is this something you can help us arrange? We have done that with other service providers.
There was a problem hiding this comment.
Hi @sschimper-splunk, thanks for the heads-up. I will see how I can help!
There was a problem hiding this comment.
Alternatively, we could perhaps mock the response with a web server and override the endpoint being passed into the splunk-observability scaler?
There was a problem hiding this comment.
The problem with that approach is that any change in the cloud service will not be detected and will require mocks update. Honestly, I prefer to not use mocks for e2e tests. Does this service run as docker image by chance?
| Realm string `keda:"name=realm, order=authParams"` | ||
| Query string `keda:"name=query, order=triggerMetadata"` | ||
| Duration int `keda:"name=duration, order=triggerMetadata"` | ||
| TargetValue float64 `keda:"name=targetValue, order=triggerMetadata"` |
|
|
||
| s.logger.V(1).Info("Started MTS stream.") | ||
|
|
||
| time.Sleep(time.Duration(s.metadata.Duration * int(time.Second))) |
There was a problem hiding this comment.
The SignalFlow Go Client library that I am using to integrate with Splunk O11y does not let you query a precise value, instead it provides a stream value series for a specified timeframe. I use this sleep to say "let the stream open for <s.metadata.Duration> seconds and then close it.
There might be some more elegant ways, but I am still fairly new to Golang and would very much appreciate feedback. An example that uses Goroutines can be found here: https://github.com/signalfx/signalflow-client-go/blob/main/signalflow/example_test.go
But I found this a bit over-kill
There was a problem hiding this comment.
Yeah, please, we should avoid using Sleep
| } | ||
|
|
||
| func (s *splunkObservabilityScaler) getQueryResult() (float64, error) { | ||
| comp, err := s.apiClient.Execute(context.Background(), &signalflow.ExecuteRequest{ |
There was a problem hiding this comment.
let's propagate the context until here.
| s.logger.V(1).Info("Started MTS stream.") | ||
|
|
||
| time.Sleep(time.Duration(s.metadata.Duration * int(time.Second))) | ||
| if err := comp.Stop(context.Background()); err != nil { |
There was a problem hiding this comment.
let's propagate the context until here.
| {&testSplunkObservabilityMetadata[0], 0, "demo-trans-latency"}, | ||
| {&testSplunkObservabilityMetadata[0], 1, "demo-trans-latency"}, |
There was a problem hiding this comment.
I think that this test is failing because the metric name must contain the index (it's done by GenerateMetricNameWithIndex)
| {&testSplunkObservabilityMetadata[0], 0, "demo-trans-latency"}, | |
| {&testSplunkObservabilityMetadata[0], 1, "demo-trans-latency"}, | |
| {&testSplunkObservabilityMetadata[0], 0, "s0-signalfx"}, | |
| {&testSplunkObservabilityMetadata[0], 1, "s1-signalfx"}, |
…names and propagate context to getQueryResult() func
Provide a description of what has been changed
Checklist
Fixes #
Relates to #