Currently the application only implements count as parameter to retrieve this amount of lines from the tail of the unit log. This makes it hard to read earlier logs, if there are too many, as they all need to go throught the LLM and eat up context and tokens, or worst case exceed its capacity.
We should have an option to select by
date: has a lower and upper limit
chunk: specifies an amount and position, to e.g. read last 100, second last 100, third last 100, or has a start point to seek forward from there
to systematically work through bigger log files. Maybe there are other/better options.