We are seeing some cases where the recent release of datadog agent with collector for crio is causing significant increase in the crio log size. It is possible to control the frequency of these check but for cases where it's absolutely needed to poll at high frequency the logs do become a problem specially due to the imageInfo object(with verbose true). (5 sec freq with 7-8pods this grows upto 40gb in a week)

Would it make sense to make the log level for https://github.com/cri-o/cri-o/blob/8a6210ef692f418ecb9c76179929d7a718e3c28e/server/image_status.go#L78C2-L78C42 to "Debug" instead of info?