changefeed resolved ts stuck under large single table scenario

[lidezhu]

What did you do?

1. create a large table with 180K regions;
2. set cdc.min-ts-interval to 200ms;

What did you expect to see?

changefeed resolved ts advance normally;

What did you see instead?

changefeed resolved ts stuck and puller memory quota are full;

Versions of the cluster

Upstream TiDB cluster version (execute SELECT tidb_version(); in a MySQL client):

(paste TiDB cluster version here)

Upstream TiKV version (execute tikv-server --version):

(paste TiKV version here)

TiCDC version (execute cdc version):

(paste TiCDC version here)