Skip to main content
Explore your complete experiment history with Tracer’s powerful search and filtering capabilities.

​
Overview

Tracer maintains a comprehensive history of all your experiments, making it easy to find past runs, compare results, and reproduce successful experiments.

​
Accessing Experiment History

1

Open History View

Navigate to the Experiments section in the Tracer dashboard.
2

Browse or Search

Use the search bar or browse through your experiment timeline.
3

Filter Results

Apply filters to narrow down your search.

​
Search Capabilities

Search experiments by:
  • Experiment Name - Find experiments by their identifier
  • Date Range - Filter by when experiments were run
  • Status - Show only successful, failed, or running experiments
  • Tags - Search by custom tags you’ve applied

​
Advanced Filters

Combine multiple filters to find exactly what you’re looking for.
Filter by:
Filter TypeDescriptionExample
DurationExecution timeTasks longer than 1 hour
Resource UsageCPU, memory, I/OExperiments using >16GB RAM
UserWho ran the experimentExperiments by team member
EnvironmentWhere it ranAWS Batch, local, Docker
ParametersInput parametersSpecific configuration values

​
Comparing Experiments

​
Side-by-Side Comparison

Select multiple experiments to compare:
  • Performance Metrics - Execution time, resource usage
  • Outputs - Result files and data
  • Configuration - Parameter differences
  • Resource Consumption - Cost and efficiency
Use the comparison view to identify which parameters led to better performance.

​
Trend Analysis

View trends across experiments: 
# View resource trends over time
tracer trends --metric memory --range 30d

# Compare success rates
tracer stats --group-by workflow --range 7d

​
Reproducing Experiments

​
Exact Reproduction

Tracer captures everything needed to reproduce an experiment:
  • Operating system and kernel version
  • Software dependencies and versions
  • Environment variables
  • Container images (if used)
  • Input file checksums
  • Data locations
  • Parameter values
  • Configuration files
  • Resource allocations
  • Execution order
  • Parallelization settings
  • Random seeds (if applicable)

​
Re-running Experiments

# Re-run an experiment by ID
tracer rerun experiment-id

# Re-run with modified parameters
tracer rerun experiment-id --param batch_size=64

# Re-run on different infrastructure
tracer rerun experiment-id --env aws-batch

​
Organizing Experiments

​
Tagging

Add tags to organize experiments: 
# Tag an experiment
tracer tag experiment-id production validated

# Search by tag
tracer search --tag production

​
Collections

Group related experiments:
  • By Project - Organize by research project
  • By Pipeline - Group by workflow type
  • By Goal - Categorize by objective

​
Exporting History

​
Export Options

Export experiment data for external analysis:

CSV Export

Export metrics and metadata to CSV

JSON Export

Full experiment details in JSON format

Report Generation

Generate PDF reports with visualizations

API Access

Programmatic access via REST API

​
Example Export

# Export last 30 days to CSV
tracer export --format csv --range 30d --output experiments.csv

# Export specific experiments
tracer export --ids exp1,exp2,exp3 --format json

​
Search Tips

Pro Tip: Save frequently used searches as “Saved Views” for quick access.

​
Keyboard Shortcuts

  • Ctrl/Cmd + K - Quick search
  • Ctrl/Cmd + F - Filter panel
  • Ctrl/Cmd + Click - Select multiple experiments

​
Search Syntax

Use advanced search syntax: 
# Search by date range
created:>2024-01-01 created:<2024-12-31

# Search by status
status:failed

# Search by resource usage
memory:>8GB cpu:>4

# Combine filters
status:success duration:>1h tag:production

​
Integration with Analysis Tools

Connect Tracer history to your analysis workflow:
  • Jupyter Notebooks - Query experiment data programmatically
  • R/Python Scripts - Analyze trends and patterns
  • BI Tools - Connect to Tableau, PowerBI, etc.
  • Custom Dashboards - Build visualizations with Grafana

​
Next Steps

Viewing Task Status

Monitor your running tasks

Investigating Task Failures

Debug and resolve failures