Job Comparison Explained🔗
The Job Comparison view provides structured performance metrics for completed scan jobs.
It allows you to compare multiple jobs side-by-side and understand:
- Overall throughput
- DSXA scan performance
- End-to-end processing cost
- Latency characteristics
- Capacity projections
This page explains what each metric represents and how to interpret it.
For step-by-step tuning guidance, see Operations → Performance Tuning.
What a Job Represents🔗
A Job is a logical scan operation initiated via:
- UI
- API
- Scheduled or automated trigger
A Job consists of many individual Scan Requests, each representing one file.
Job metrics aggregate performance across all files processed in that scan.
Core Throughput Metrics🔗
Job Time🔗
Total elapsed time from job start to completion.
This is the primary end-to-end performance indicator.
It includes:
- Enumeration
- File retrieval
- DSXA scanning
- Verdict handling
- Result persistence
Total Bytes🔗
The total size of all files processed in the job.
This represents workload volume.
Total Bytes/sec🔗
End-to-end throughput across the entire system.
Formula (conceptually):
id="v7fl6h"
Total Bytes ÷ Job Time
This includes:
- Connector I/O
- Network transfer
- DSXA scan time
- Queue overhead
- Remediation time
This metric reflects overall system performance.
Scan Bytes/sec🔗
Throughput inside DSXA only.
This excludes:
- Enumeration time
- File retrieval latency
- Connector overhead
If Scan Bytes/sec is stable but Total Bytes/sec drops, the bottleneck is likely outside DSXA.
Latency Metrics🔗
Avg Req ms🔗
Average time required to process a single file request.
This includes:
- Connector read time
- Network transfer
- DSXA processing
- Verdict handling
Higher values may indicate:
- Large files
- Slow connectors
- Network latency
- DSXA saturation
Scan ms/byte🔗
Average scanning cost per byte inside DSXA.
This allows comparison across different file size distributions.
Useful for estimating how scan performance scales with larger datasets.
Progress Metrics🔗
Processed🔗
Number of items processed versus total items discovered.
Indicates job completion progress.
Status🔗
Indicates whether the job completed successfully, failed, or was interrupted.
Projection Metrics🔗
The Job Comparison view provides normalized estimates such as:
- Estimated Job Time: 1GB
- Estimated Job Time: 1TB
- Estimated Job Time: 1M files
- Estimated DSXA Time: 1GB
- Estimated DSXA Time: 1TB
These projections are derived from:
- Observed throughput
- Observed file size distribution
- Observed scan cost per byte
They allow you to answer questions like:
- “How long would 5TB take?”
- “What happens if this dataset doubles?”
- “Is this throughput acceptable for my SLA?”
Understanding DSXA Time vs Job Time🔗
Two different projections are provided:
- Estimated DSXA Time — time spent inside scanning only.
- Estimated Job Time — full end-to-end processing time.
If DSXA Time is much lower than Job Time:
- Enumeration or connector I/O likely dominates.
If DSXA Time closely matches Job Time:
- Scanning itself is the primary bottleneck.
Interpreting Differences Between Jobs🔗
When comparing jobs:
Look for changes in:
- Total Bytes/sec
- Scan Bytes/sec
- Avg Req ms
- Estimated projections
Differences may indicate:
- Infrastructure saturation
- Different file mixes
- Changes in concurrency or replicas
- Changes in connector configuration
- Changes in DSXA capacity
The Job Comparison view does not prescribe tuning steps. It provides measurement data for informed decisions.
Important Notes on Projections🔗
Projections are estimates based on the scanned dataset.
Accuracy depends on:
- File size distribution
- File type mix
- Network conditions
- DSXA characteristics
Always validate projections using representative datasets.
Measurement Before Modification🔗
The purpose of Job Comparison is measurement.
It allows you to:
- Establish a baseline
- Compare configuration changes
- Quantify performance gains
- Evaluate cost/performance tradeoffs