Improves Redis cache list retrieval performance #116
Conversation
Optimizes list retrieval by using SortedSetRangeByScoreAsync instead of SortedSetRangeByRankAsync, utilizing timestamp-based scores for efficient expiration management. Handles legacy sets by migrating them to sorted sets when encountering WRONGTYPE exceptions, ensuring compatibility and preventing data loss during the transition. Logs trace information when a sorted set is empty during expiration setting.
There was a problem hiding this comment.
Pull Request Overview
This PR optimizes Redis cache list retrieval performance by switching from rank-based to score-based queries, using timestamps as scores for efficient expiration management. It also adds backward compatibility by migrating legacy sets to sorted sets when WRONGTYPE errors occur.
Key Changes:
- Replaced
SortedSetRangeByRankAsyncwithSortedSetRangeByScoreAsyncfor improved query performance - Added exception handling to migrate legacy sets to sorted sets automatically
- Extracted migration logic into a separate reusable method
| return RedisValuesToCacheValue<T>(set); | ||
| if (!page.HasValue) | ||
| { | ||
| var set = await Database.SortedSetRangeByScoreAsync(key, _timeProvider.GetUtcNow().ToUnixTimeMilliseconds() + 1, flags: _options.ReadMode).AnyContext(); |
There was a problem hiding this comment.
The magic number '+1' added to the timestamp lacks explanation. Consider adding a comment explaining why this offset is necessary (e.g., to include items with scores equal to the current timestamp).
| else | ||
| { | ||
| long skip = (page.Value - 1) * pageSize; | ||
| var set = await Database.SortedSetRangeByScoreAsync(key, _timeProvider.GetUtcNow().ToUnixTimeMilliseconds() + 1, skip: skip, take: pageSize, flags: _options.ReadMode).AnyContext(); |
There was a problem hiding this comment.
The magic number '+1' added to the timestamp lacks explanation. Consider adding a comment explaining why this offset is necessary for consistency with the non-paginated query.
Updates the sorted set range query for retrieving cached lists to use Double.PositiveInfinity for the maximum score, ensuring all valid items are returned. This change enhances performance by avoiding unnecessary additions of 1 millisecond to the timestamp, streamlining the retrieval process and improving efficiency.
![github-code-quality[bot]](https://avatars.githubusercontent.com/u/9919?s=60&v=4){kind=small}
| catch (Exception ex) | ||
| { | ||
| _logger.LogError(ex, "Unable to delete {HashSlot} keys ({Keys}): {Message}", hashSlotGroup.Key, hashSlotKeys, ex.Message); | ||
| } |
Optimizes list retrieval by using SortedSetRangeByScoreAsync instead of SortedSetRangeByRankAsync, utilizing timestamp-based scores for efficient expiration management.
Handles legacy sets by migrating them to sorted sets when encountering WRONGTYPE exceptions, ensuring compatibility and preventing data loss during the transition.
Logs trace information when a sorted set is empty during expiration setting.