[mempool] reduce default capacity per user

[bchocho]

Description

Reduce the default capacity per user: 100 -> 20. This should make it less likely that some users take over mempool capacity.

Test Plan

Ran land_blocking and load_vs_perf_benchmark with good results. Will let CICD run the continuous tests.

---

This change is 

[bchocho]

The compat failures seem like an unrelated change in main. The land blocking failures are legit. Will have to investigate.

[igor-aptos]

Revert main.rs before landing

[bchocho]

This is what I got from load_vs_perf before it timed out (due to setting too short of a duration)

wanted/s     | submitted/s  | committed/s  | expired/s    | rejected/s   | chain txn/s  | latency      | p50 lat      | p90 lat      | p99 lat      | actual dur  
7500         | 7502         | 7060         | 442          | 0            | 6645         | 52126        | 60100        | 69500        | 79300        | 720         
8000         | 8003         | 6700         | 1303         | 0            | 6486         | 62638        | 64300        | 69400        | 79100        | 720         
9000         | 8999         | 6256         | 2742         | 0            | 6281         | 65866        | 64500        | 70300        | 80100        | 720         
10000        | 10004        | 6260         | 3743         | 0            | 6201         | 64635        | 63700        | 68000        | 73300        | 720         
12000        | 11993        | 6187         | 5805         | 0            | 6170         | 64522        | 63300        | 68500        | 77100        | 720         

I also noticed account minting takes longer (7m vs 1m previously) likely because of the parallelism reduction.

[sitalkedia]

Are the failures in Forge tests unrelated to this change? Can we get a successful run?

[bchocho]

I had set land_blocking to load_vs_perf_benchmark and didn’t set the duration long enough so it timed out. Github should run one more land_blocking before land :)

[github-actions]

✅ Forge suite compat success on 843b204dce971d98449b82624f4f684c7a18b991 ==> 2af5876cd74b95bd0db4182bc3eb118f7a26f962

Compatibility test results for 843b204dce971d98449b82624f4f684c7a18b991 ==> 2af5876cd74b95bd0db4182bc3eb118f7a26f962 (PR)
1. Check liveness of validators at old version: 843b204dce971d98449b82624f4f684c7a18b991
compatibility::simple-validator-upgrade::liveness-check : 7748 TPS, 4887 ms latency, 8800 ms p99 latency,no expired txns
2. Upgrading first Validator to new version: 2af5876cd74b95bd0db4182bc3eb118f7a26f962
compatibility::simple-validator-upgrade::single-validator-upgrade : 5211 TPS, 7396 ms latency, 10000 ms p99 latency,no expired txns
3. Upgrading rest of first batch to new version: 2af5876cd74b95bd0db4182bc3eb118f7a26f962
compatibility::simple-validator-upgrade::half-validator-upgrade : 4735 TPS, 8056 ms latency, 12800 ms p99 latency,no expired txns
4. upgrading second batch to new version: 2af5876cd74b95bd0db4182bc3eb118f7a26f962
compatibility::simple-validator-upgrade::rest-validator-upgrade : 7850 TPS, 4757 ms latency, 8100 ms p99 latency,no expired txns
5. check swarm health
Compatibility test for 843b204dce971d98449b82624f4f684c7a18b991 ==> 2af5876cd74b95bd0db4182bc3eb118f7a26f962 passed
Test Ok

• Grafana dashboard
• Humio Logs
• Test runner output
• Test run is land-blocking

[github-actions]

✅ Forge suite land_blocking success on 2af5876cd74b95bd0db4182bc3eb118f7a26f962

performance benchmark with full nodes : 7579 TPS, 4850 ms latency, 7200 ms p99 latency,(!) expired 20074 out of 3256500 txns
Test Ok

• Grafana dashboard
• Humio Logs
• Test runner output
• Test run is land-blocking