CrashLoopBackOff means the container keeps exiting and the kubelet keeps restarting it, waiting longer each time. The question is always why it exits, and the evidence tells you.
Start with kubectl describe pod. Last State shows how the previous run ended: exit code 1 (or another small number) means the application stopped by itself; 137 means it was killed, by a failed liveness probe (look for Killing events) or by its memory limit (Reason: OOMKilled). Then read the crashed run's output with kubectl logs POD --previous; the current container has only just started.
Applications usually crash at start for a missing or wrong setting, an unreachable dependency, or a bad command. Fix the input, not the symptom: restarting re-runs the same broken template.
Init:0/1 is a different stall: an init container hasn't finished. Read its logs with -c to see what it waits for.
Status, Last State and events, the crashed run's output, an init container.