Can you understand what’s happening in your system and why, without having to push new code?

Maybe, maybe not.

But do you know what’s really hard to do if you can’t?

Add meaningful o11y after the fact because you don’t get value from history.

This goes past “do you have logs” and “do you have metrics” because that’s, just, what it is.

We know the quickest and best way to get validation from tests is to be able to run them locally, can you look at your structured logs and your events locally to check if they are providing you value?

If you aren’t, consider looking at something like grafana/otel-lgtm which will stand up an OTEL stack with Grafana, Tempo for traces, Loki for logs and Prometheus for metrics.

If you’re using .NET I freaking love the look of .NET Aspire to get this feedback simply and out of the box as a local development framework.

Between those you can get feedback loops early in your development processes to make sure that you’re logging properly, that your metrics are answering meaningful.