Not getting right number of revisions on SGW while using multiple SGWS

CBL:  2.7.0-30 - all mercury builds

Steps to Reproduce:
1. Create 2 DBs in 2 SGS and Create docs in CBL
2. Set up 1 SG with revs limit 30 and SG2 with revs limit 25.
3. Do push replication to two SGS(each DB
to each SG).
4. Do updates on CBL for 35 times.
5. Continue push replication to SGs
6. Verify that revs maintained on two SGS according to revs_limit.

• Actual Result:

Revision limit shows the range of 20 to 30 on first DB of SGW. Most of the time revisions are not maintained until 30 , but I see rev generation is 36 on SGW which shows all 35 updates are pushed to SGW.

All the updates which are updated on CBL are pushed to SGW successfully , just that revisions are not maintained

• Expected Result:

All 30 revisions should exists in SGW as revision limit for DB1 is 30 

CBL LOGS: https://gist.github.com/sridevi-15/057643a525ad1b0dc0406e83213540b7
Github link for the code:

     https://gist.github.com/sridevi-15/057643a525ad1b0dc0406e83213540b7
Jenkins job failure link:

http://uberjenkins.sc.couchbase.com:8080/job/CBLITE_iOS-Listener-TestServer-Topology-specific-tests/898/testReport/testsuites.CBLTester.topology_specific_tests.multiple_sync_gateways/test_replication_multiple_sgs/test_multiple_sgs_with_differrent_revs_limit_listener_tests_multiple_sync_gateways_100_/
Pytest Command
pytest --timeout 1800 --liteserv-version=2.7.0-30 --liteserv-host=10.100.172.194 --liteserv-port=8080 --no-conflicts --device --enable-file-logging --delta-sync --sync-gateway-version=2.7.0-24 --mode=cc --server-version=6.5.0-3939 --liteserv-platform=ios --create-db-per-test=cbl-test testsuites/CBLTester/topology_specific_tests/multiple_sync_gateways -k test_multiple_sgs_with_differrent_revs_limit[listener_tests/multiple_sync_gateways-100]

What is the last build this test passed: Cobalt release builds of CBL

Failing only in mercury builds

I tested with SGW cobalt and release builds and it is still failing. 

Happening on all 3 platforms : iOS, Android, .Net

We have tests with number of docs 10, 100, 1000.

Couple of time, number of docs with 10 fail, but most of the time in Mercury is getting failed with number of docs 10, 100, 1000