On the search route, we return the latency of the query for each component in the timing headers and we should display the sum of these:  The end result should look similar to what Algolia does: 