Is it possible to scale when threshold drops below a value?

[jesumyip]

I have a situation where a particular task is run on a schedule (once every 2seconds). This tells me if the process running the task is still alive. I export this to Newrelic

If the process is overwhelmed, that task may only run at the next 15second window instead of 2 second.

I want keda to autoscale my pods if the NRQL query returns a value of less than 30 for a SELECT count. Most of the examples I have seen is for keda to trigger autoscaling when the value is MORE than 30.

[jesumyip]

https://keda.sh/docs/2.12/concepts/scaling-deployments/#scaling-modifiers-experimental

It looks like the ternary operator would be what I require?

[zroubalik]

Yeah, definitely give it a try and please report back if you were successful.

[jesumyip]

Follow-up question. Is there a way I can get the current number of replicas to be used in calculations? Example:

1. I have a readinessProbe that runs every 5 seconds.
2. Each replica therefore generate 12 readinessProbe calls (assuming all are successful) per minute.
3. In NewRelic, I can see 120 probes generated every 10minutes if I have 1 replica.
4. If Keda scales this deployment up to 2 replicas, that NRQL then changes to 240 probes per 10minutes.
   How can I update automatically the Keda ScaledObject to now watch for 240 probes instead of 120 probes per 10 minutes?