Explain the MongoDB aggregation pipeline.

Aggregation pipeline processes documents through stages: $match (filter), $project (reshape), $group (aggregate by key), $sort, $limit, $skip, $unwind (deconstruct arrays), $lookup (join collections), $facet (multiple pipelines). Documents flow through stages in order. Each stage transforms documents and passes to next. Example: match active users, group by country, calculate average age. Similar to SQL GROUP BY but more flexible. Consider indexes on $match fields for performance.