Part 5
Running a Healthcare Claims Platform Locally in Kubernetes, Part 5: From Faster to Repeatably Faster
How the local Kubernetes sweep harness turned isolated fast runs into repeatable performance evidence.
Engineering series
Million Claim Challenge Field Notes
How the benchmark moved from local Kubernetes runs to repeatable measurement, honest workflow scoring, and operator-facing evidence.
Current verified scope: the latest asynchronous local Kubernetes run reached the full 1,000,000-claim corpus at 155.89 claims/sec, with all claims eventually terminal. A separate 100,000-claim raw X12 837 run reached 199.42 claims/sec end-to-end. These are local validation results, not production-cloud capacity claims.
Part 6
Running a Healthcare Claims Platform Locally in Kubernetes, Part 6: From Fast Runs to Honest Edge-Case Scoring
Why paid is not the only correct outcome, and why unsupported scenarios must remain visible.
Part 7
Running a Healthcare Claims Platform Locally in Kubernetes, Part 7: From Benchmark Logs to an Operator Console
Moving benchmark proof from terminal logs into an operator-facing Mass Adjudication console.
Part 8
Running a Healthcare Claims Platform Locally in Kubernetes, Part 8: The Clean 100,000-Claim Run
How stronger correctness gates reached a clean 100,000-claim run and exposed the next local scaling bottleneck.
Part 9
Running a Healthcare Claims Platform Locally in Kubernetes, Part 9: From Unsupported to Scored
How prior-auth wrong-provider and behavioral-health scenarios moved from unsupported to deliberately scored platform behavior.
Part 10
Running a Healthcare Claims Platform Locally in Kubernetes, Part 10: The Migration Cost That Wasn't
Testing Part 9's unconfirmed migration-cost theory at 250,000 claims, finding the real causes through profiling, and confirming a 3.5x throughput gain.
Part 11
Running a Healthcare Claims Platform Locally in Kubernetes, Part 11: The Check That Only Ran in the Benchmark
Re-investigating a bug Part 10 disclosed but didn't fix, and finding federal provider-exclusion screening had never been wired into the real adjudication pipeline at all.
Part 12
Running a Healthcare Claims Platform Locally in Kubernetes, Part 12: The Database Nobody Profiled
Two benchmark fixture bugs, a Submit-chain bottleneck traced to an under-provisioned shared database, and this series' first clean 500,000-claim confirmation.
Part 13
Running a Healthcare Claims Platform Locally in Kubernetes, Part 13: The Gap Was the Laptop
Closing the wall-clock gap disclosed in Part 10 and Part 12 by proving, with matching host sleep-log evidence, that it was macOS suspending the local Kubernetes cluster mid-run.
Part 14
Running a Healthcare Claims Platform Locally in Kubernetes, Part 14: Zero Unsupported, Then the Parallelism Nobody Profiled
Closing the scoring gap carried since Part 9 with a first-ever zero-unsupported run, then finding and fixing why parallelism 56 had quietly underperformed lower concurrency this whole series.
Part 15
Running a Healthcare Claims Platform Locally in Kubernetes, Part 15: One Million Claims, and the Bug Only That Scale Could Find
Running this series' first full 1,000,000-claim confirmation, finding a Redis memory ceiling that only that scale could expose, and fixing it live.
Part 16
Running a Healthcare Claims Platform Locally in Kubernetes, Part 16: The Million Went Through the Bus
Moving the full million-claim corpus onto asynchronous Service Bus adjudication, then proving the separate raw X12 837 onramp at 100,000 claims.