R&D advanced 3 min read May 4, 2026
Public Preview Sign in free for the full digest →

Wise: 1,000+ Services, One Chassis, £36B Quarterly

“A single versioned Gradle artifact governs security, observability, and data access across all 1,000+ Wise services — and a bug in it could cascade to every one of them simultaneously.”

Wise: 1,000+ Services, One Chassis, £36B Quarterly
2 Views
0 Likes
0 Bookmarks
Source · blog.bytebytego.com

“"Make the right engineering choice the easy one." — Wise platform engineering philosophy (source: ByteByteGo article citing Wise Engineering Blog, verified 2026-05-04)”

You know that feeling when your 50th microservice has slightly different logging than your 1st, your 30th service has a custom authentication setup three versions behind, and your security team can't audit any of it because nothing is consistent? At 850 engineers and 1,000+ services, Wise faced exactly this: how do you stop every team from independently reinventing security middleware, observability config, and database connection setup — and then diverging from each other over 18 months? The naive answer is documentation and code reviews. The real answer is making the standard the only dependency in the build file.

microservicesplatform-engineeringfintechjavakubernetessystem-designkafka

Wise's backend runs 1,000+ Java/Kotlin microservices, all consuming a single versioned Gradle artifact called the Microservice Chassis. Think of it like a standard electrical socket: every appliance (service) plugs in and immediately gets power, grounding, and surge protection (security, observability, DB connections, Kafka) — no team configures these from scratch. When Wise's platform team bumps the chassis version, all 1,000+ services inherit the change at their next build. Deployments flow through GitHub Actions CI into Spinnaker, which routes 5% of real traffic to the new version for a 30-minute observation window before full rollout — with automatic rollback if anomalies appear. The entire observability layer (Loki for logs, Grafana for dashboards, Tempo for traces, Mimir for metrics) runs on dedicated isolated clusters, ingesting 6 million metric samples per second without competing for production compute.

01
Microservice Chassis (versioned Gradle artifact) — instead of auditing 700+ repos for security and observability drift, you bump one version number and all 1,000+ services inherit the change at their next build without per-team action
02
Spinnaker canary deployments (5% traffic, 30-min window, auto-rollback) — per Wise's own 2024 self-reported data, this configuration stopped hundreds of incident-causing deployments before they hit full traffic
03
LGTM observability stack on isolated clusters — Loki, Grafana, Tempo, and Mimir run on dedicated compute separate from production, so a traffic spike cannot take down your monitoring at the exact moment you need it
04
GitHub Actions CI with pre-populated container caches — migrating from CircleCI gave 15% build speed improvement; pre-populated caches added another 15%; net result is 1,000+ engineer-hours saved per month across 500,000 builds
05
iOS build time: 28s → 2s via Swift Package Manager migration — replacing Cocoapods with Tuist and SPM across 250+ Xcode modules cut build time by 93%, directly freeing iOS engineers from waiting
06
Multi-provider LLM gateway — routes requests across Anthropic Claude, AWS Bedrock, Google Gemini, and OpenAI from a single internal Python library so swapping or A/B-testing models requires no application code changes
07
Trino federated queries — engineers query across Snowflake, Apache Iceberg on S3, and live Kafka streams from one SQL interface, without ETL pipelines to move data between systems first
Who it’s for

If you are a platform engineer, backend architect, or engineering manager at a company with 50+ engineers and more than 20 microservices, this stack gives you a concrete reference point for platform standardization. The Microservice Chassis pattern is especially relevant if your services are visibly diverging on security and observability config. Not useful if you are pre-product-market-fit or running fewer than 10 services — the platform team investment required to maintain a shared versioned artifact is only justified when service count pushes coordination overhead past what code reviews ca...

Worth exploring

Yes, if you are designing a platform engineering strategy for a 50+ engineer org. The Microservice Chassis pattern, LGTM observability stack, and Spinnaker canary approach are all either open-source or reproducible with open-source tooling. The caveat: every number in this article is self-reported by Wise's engineering team with no independent verification — the original Wise Engineering Blog post received 1 HN point and 0 comments, so none of these claims have been publicly challenged or corroborated.

Developer playbook
Tech stack, code snippet, sentiment, alternatives.
PM playbook
Adoption angles, user fit, positioning.
CEO playbook
Traction signals, ROI, build vs buy.
Deep-dive insight
Full long-form analysis, no fluff.
Easy mode
Core idea, fast — when you need the gist.
Pro mode
Technical nuance, edge cases, tradeoffs.
Read the full digest
Go beyond the preview

Deep-dive insight, Easy and Pro modes, plus action playbooks — the full breakdown is one tap away.

Underrated tools. Unfiltered takes.

Read the full digest in the Snaplyze app for deep-dive insight, Easy and Pro modes, and the playbooks you can actually use.

Install Snaplyze →