optimize loading of time entries

[HenrikeW]

Description

• Check how exactly time entries for a week are loaded right now
• see if there is a better way of doing it, e.g. fetching all time entries for the whole week in one query and then filter what we need.

Info

If this seems difficult or takes a long time, do #906 first.

[kusalananda]

Currently, the frontend code appears to make one Redmine query per row. These are essentially variations of the following query, with different combinations of issue_id and activity_id values:

/time_entries.json?user_id=me&issue_id=6818&activity_id=13&from=2024-10-14&to=2024-10-18&offset=0&limit=100&user_id=me

These could quite likely be combined into a single query, by removing the issue_id and activity_id,

/time_entries.json?user_id=me&from=2024-10-14&to=2024-10-18&offset=0&limit=100&user_id=me

... that the code later parses the needed information out of. Care must be taken to ensure that all information is fetched, either by increasing the limit parameter to some larger number, but ideally by implementing proper paging (multiple queries with an increasing offset value to get further data), or a combination thereof.

Paging like this is already implemented for computing the "Yearly overview". To show this bar all time entries for the whole year are fetched (which is a possible reason users start complaining about slowness towards the end of the year (!)).

This is an example of a query made to construct the "Yearly overview" bar:

/time_entries.json?user_id=me&from=2024-01-01&to=2024-12-31&offset=500&limit=100&user_id=me

[kusalananda]

There are also some clear cases for optimising the backend Go code, especially the code in getPriorityEntriesHandler.go. I'm adding the "backend" label to this issue. This likely means that this issue will have two PRs.