Kafka sidecar: add a startupProbe so slow-starting functions aren't restart-looped #4097

What

The injected Kafka runtime sidecar (kafka-runtime container, added by the
raw/keda deployers in pkg/k8s/deployer.go) has a liveness probe but no
startupProbe (nor an initialDelaySeconds sized for cold start).

Problem

The runtime only begins serving /health/liveness after it finishes waiting for
the function target (an up-to-~60s FUNCTION_TARGET wait). A slow-starting
function — e.g. a JVM cold start that exceeds the liveness threshold — can fail
liveness before it is ever up, so the kubelet restart-loops the sidecar (and the
pod never becomes healthy).

Fix

Add a startupProbe to the sidecar container (or an initialDelaySeconds that
covers the worst-case cold start) so the liveness probe only takes effect after
startup completes. Readiness and liveness themselves landed in #4060; this closes
the remaining startup-probe gap.

/kind bug

Follow-up to #4060.