Client: CargoMatch · Logistics, Travel & Marketplaces · EU freight marketplace (70 staff)

Reducing Failed Matches on a Freight Marketplace

By TechSupport Team. Published November 16, 2026. Updated .

TechSupport Team supported CargoMatch’s Python and Redis freight matching service. Over four months, client Datadog records show failed matches falling from 8.9% to 2.6%. P95 matching time fell from 2.4 seconds to 610 ms.

Client and engagement facts

Client
CargoMatch
Industry
Logistics, Travel & Marketplaces
Business context
EU freight marketplace (70 staff)
Delivery model
Application support
Application
Freight matching platform with load dispatch
Supplied geography
EU
Company size in source
70 staff

Application stack and tools

Application stack

  • Python
  • PostgreSQL
  • Redis
  • AWS

Tools used

  • Datadog
  • Sentry

Why Support Was Needed

The team found three contributors to failed matches. Redis keys expired at the wrong time, the matching algorithm had a race condition and the fallback matcher missed edge cases.

Technical Work

  • Fixed Redis key expiry causing 60% of failures.
  • Fixed the matching algorithm race condition.
  • Added a fallback matcher for edge cases.

Representative Incident

A driver app version mismatch caused a one-day failure spike. TechSupport Team documented the cause for the client’s mobile team, which owned that component.

Results and how they were measured

Reported results and their measurement scope
MetricBaselineResultUnitsMeasurement windowSource and scope
Failed matches8.9%2.6%%4 monthsClient Datadog
Matching P952.4 s610 msBaseline: s; Result: ms4 monthsClient Datadog
App crash rate3.1%0.4%%4 monthsClient Sentry

Evidence and Limitations

The four-month results exclude one driver-app version mismatch. Matching measures rely on Datadog, and app crash rate comes from Sentry.

Matching Reliability Review

TechSupport Team added a matching success dashboard to CargoMatch’s Datadog account. Both teams reviewed it weekly.

Scope a Similar Engagement

Describe the matching or dispatch errors you see, including which parts are owned by another team. Request a support plan or explore application support.

Get a technical support plan

Tell us which application needs support, what is reaching your engineers and when you need coverage. TechSupport Team will review the fit and explain the next scoping step.

  • We reply by email.
  • We agree a call about the roles, your stack and the hours.
  • If there is a fit, we propose engineers after an assessment.

A short outline is enough. Add your stack and coverage hours if you know them. Do not include credentials or customer data.

Add coverage and team details

We use these details only to reply to you. Describe the application, stack and hours you need covered. Keep credentials and sensitive technical data out of this form. We save the brief for delivery and use Telegram to notify our team.