The Challenge
A global logistics provider used search within an application that helped employees route shipments throughout its international network. Individual search nodes periodically crashed. Redundancy allowed operations to continue while affected nodes were restarted, but the organization needed greater search reliability as the application expanded into increasingly critical freight-management workflows.
The client engaged Innovent for specialist search expertise to assess the implementation and improve its architecture and configuration. The work needed to account for frequent data updates, availability requirements, and the operational consequences of search failures.
Innovent’s Approach
Innovent examined the deployment topology, collection sizes, runtime configuration, query behavior, and caching. The assessment found that relatively small collections were distributed across multiple shards, adding coordination and query complexity without a corresponding capacity requirement.
Innovent simplified the search architecture to use a single shard with multiple replicas, retaining redundancy while removing unnecessary distribution. The team also separated cluster coordination services from the search servers and moved the search engine from a shared application-container arrangement to a dedicated runtime, enabling clearer resource isolation and tuning.
Another finding concerned the relationship between data freshness and cache management. The application made updates searchable every 2.5 seconds to support employees entering shipment and routing information. Cache warming repeatedly replayed queries against newly updated indexes, consuming resources even though the caches showed little evidence of useful activity.
Innovent revised the cache configuration and warming approach to better suit this frequent-update workload. The changes addressed recurring processing overhead while preserving the application’s need for current information.
The assessment also covered hardware resources, runtime tuning, security, monitoring, and deployment practices. Recommendations addressed the operational requirements of the shipment-routing application and the platform’s expanding business role.
The Result
The architecture and configuration changes were implemented in production. The client gained a simpler search topology with replicated availability, separately hosted coordination services, and cache behavior better aligned with frequent data updates.
Innovent strengthened the search foundation supporting shipment operations and provided guidance for maintaining and scaling the platform as its business role expanded. The work matched architectural complexity and resource use to the workload and reliability requirements of an operational application.