Fix match analysis on API usage

Problem

After the translate5 - 5.8.5 update and the introduction of pausing of match analysis in case the memory is not available, some additional improvements and fixes are required for translate5 api usage.

• currently if the analysis pause worker exceeds the waiting time and there are still unavailable memories, no exception is thrown and the analysis will continue.
• when pretranslation or analysis operation is triggered the task should be locked with analysis state immediately and not in the analysis worker. With this, on api usage, the user will get the analysis task status with the initial request.
• confluence documentation is missing for analysis pause

Developer note: check/test the translate5 frontend after those changes. Especially after the status change!