Amazon, 2025
Showing service owners where their traffic actually went
A CLI command that answers “where did my endpoint resolve last week?” Troubleshooting went from hours to minutes for more than 200 users.
Local Region Affinity (LRA) optimizes how requests are routed across Amazon's regions. It works well, but it adds a layer of abstraction: service owners could no longer easily see how their endpoints were being resolved at runtime. When something went wrong, they had no way to look back and check.
Who it was for
- Amazon service owners and engineers responsible for LRA-enabled endpoints
- Strategic customers onboarding to LRA who needed operational transparency
What users told us
I went through feedback from senior engineers, technical risk assessments, and onboarding sessions with strategic customers. The same questions kept coming up:
Where did my endpoint resolve last week? Did traffic go to the right region? Why did a request fail?
What they wanted was historical, queryable data: something they could use to investigate incidents, verify routing, and build confidence in how LRA was treating their traffic.
What I built
I designed and built resolution-history, a new command in the LRA Mechanic CLI. Service owners can query how their endpoints resolved over time, filter by timeframe, caller region, and AWS account, see summary statistics, and export the results.
- Up to one week of history
- Data refreshed every five minutes
- Fast queries on both small and large time windows
I put it in the CLI engineers already used rather than building a new tool, and backed it with Amazon Athena reading from S3, cataloged by Glue. That kept queries serverless and on-demand, with almost no operational overhead for the team.
Example
mechanic execute lra health resolution-history \
--endpoint=service.lra.amazon \
--start-time='2025-06-01T00:00:00' \
--end-time='2025-06-01T02:00:00' \
--regions=dub,zaz --account=123456789012| Region | Total calls | Successful | Failed |
|---|---|---|---|
| eu-west-1 | 50 (50%) | 48 (96%) | 2 (4%) |
| eu-south-2 | 50 (50%) | 47 (94%) | 3 (6%) |
| Total | 100 | 95 (95%) | 5 (5%) |
Sample output.
Impact
Troubleshooting regional routing went from hours to minutes for more than 200 users. Service owners can now verify resolution behavior across regions and time periods, investigate incidents with real data, and catch problems faster.
The resolution-history command has significantly improved our ability to monitor and troubleshoot our LRA endpoints. It's become an essential tool in our operational toolkit.
What's next
- Comparing resolution patterns across time periods
- Integrating with monitoring and alerting
- Visualizing long-term trends