Surprise-Free Workload Management

Workload management is one of the core and critical processes in any IT setup. Any delay or failure in your workloads can cause severe impacts to the Brand image, in addition to the financial losses it can cause. It is imperative to have a mechanism to predict the performance of the workload system and to proactively apply fixes even before the problems occur.

This process requires a cognitive approach involving a technology-agnostic comprehensive blueprint of the job streams, profiling the normal behavior analysis coupled with context-aware self-triaging and a self-healing mechanism.

The problem with existing workload management systems is that they do not consider the batch data in conjunction with the various KPIs in the infrastructure based on a context-aware system.

This post will show how Digitate’s ignio™ predicts the workload run time for surprise-free workloads.

Figure 1: Pain areas and Solutions

Unpredictable to Predictable

ignio classifies batches across a variety of buckets according to the workload, run time, throughput, criticality, and scheduling dependencies.

Figure 2: Increasing trend of a Job Workload
Figure 3: Increasing trend of Job Run Time
Figure 4: Decreasing trend of Job Throughput

From the above metrics, in tandem with the analysis of the metrics of the underlying infrastructure, ignio can predict the run time of the batch for the day and identify the levers which can be adjusted to improve the performance of the job.

ignio alerts the batch operators, giving a sufficient lead time to allow them to react and bring things under control.

Figure 5: A sample prediction mail

ignio uses its incident management capability to determine the potential issues that are causing the drag on the batch stream, achieving a surprise-free workload management system and avoiding batch firefighting.

Figure 6: ignio provides a deeper view of your system

by Karthik Subramaniam, ignio solution architect

Leave a Reply