Worked on the dragonflydb/dragonfly repository to enhance cluster coordination by introducing configurable timeout flags for connections and responses across shards. The solution replaced hard-coded timeout values in the CrossShardClient with dynamically tunable flags, allowing operators to adjust timeouts without code changes or redeployment. This approach improved both the reliability and flexibility of remote shard interactions, addressing the need for environment-specific tuning. The work was implemented in C++ and leveraged skills in network and system programming to deliver a focused feature that increased operational control. No bugs were fixed during this period, with efforts concentrated on this targeted configurability improvement.