Historic Data in Kubernetes

We usually only tail new data from logs. For very fast logs of newly launched pods, we may miss the first few kb of logs. Our agent typically only gets data from head of the file if the file itself is smaller than 8kb during its rescan interval of 1 minute. So basically if a newly launched pod generates more than 8kb of data in its first minute, we may only get new data into that pod after the rescan is complete.