Indeed operates enterprise-wide global clusters in the cloud across multiple availability zones worldwide, utilizing MongoDB for approximately 50% of its operations. Running these clusters can introduce performance issues due to initial node syncs, replication lags, and changes in sync sources. To address these challenges, Indeed uses techniques such as overriding sync sources, speeding up replication with WiredTiger cache size adjustments, and monitoring write performance to identify potential problems before they impact users. Additionally, the company configures write concerns across Availability Zones in AWS to ensure seamless data replication and minimize downtime. By adopting proactive strategies, Indeed has improved its overall resilience and performance, enabling faster response times and better end-user experiences.