We've recently updated our Elasticsearch cluster version from 6.7 to 7.8. Since we upgraded, we've noticed that the performance of our percolate queries has significantly degraded.
For reference, when running a single-threaded application in ES 6.7 that only sends percolate queries one at a time, we were able to send 37 percolate requests in a 2-minute period, and the p99 of the response time was 3.4 seconds.
By contrast, on an identical cluster using ES 7.8, and using the same percolate queries in the same order, we were only able to send 17 queries in a 2-minute period, and the p99 was 8.2 seconds.
We also tried using ES 7.9 and had similar results.
I was wondering if there are any known issues regarding percolate performance on ES 7.8 (I couldn't find anything in the release notes), and if so, is there anything we can do to mitigate the impact of this?