Quickly locate rogue jobs and storage bottlenecks and keep track of what’s running on clusters day-to-day with HPCWorks Mistral.
Monitor I/O and collect system telemetry to optimize application performance on-premises, in the cloud and in hybrid environments. Take advantage of per-job monitoring and detailed reporting to track I/O, metadata, CPU and memory for easy troubleshooting and problem resolution.
Plan for tomorrow, control costs and optimize day-to-day performance using operational telemetry and business intelligence with a lightweight, storage-and scheduler agnostic design to track real-time needs and long-term trends.
Quickly identify problems using detailed storage, I/O and application data. Fix problems and optimize performance by tracking down rogue jobs and bad I/O patterns such as excess metadata, random I/O and small reads and writes.
Use detailed data to understand how to maximize on premises, cloud or hybrid environment. Log data for each job and mount point, and view information per job, user or group. Measure CPU, memory, metadata and I/O performance.
Evaluate how applications access data to optimize performance, portability and planning. Understand I/O patterns for debugging, resource sharing and tuning, and map application dependencies for migration and containerization.
Make better decisions about cloud scheduling and deployment by forming a common language across local and cloud compute environments. HPCWorks Mistral is system- and storage-agnostic.
Integrate HPCWorks Mistral with Siemens or third-party workload managers and existing monitoring data pipelines, and push data to a wide range of databases and messaging systems.
