#quarkus — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #quarkus, aggregated by home.social.
-
JDK 25 gives Quarkus services another startup option: train an AOT cache, keep the JVM, and reuse class-loading, linking, and method-profile data on the next run.
I built the full path with PostgreSQL and Podman: packaged training test, strict cache verification, a 20-run benchmark, and a target-specific container image. Local median readiness dropped from 1022.4 ms to 380.3 ms.
https://www.the-main-thread.com/p/quarkus-project-leyden-aot-cache
-
JDK 25 gives Quarkus services another startup option: train an AOT cache, keep the JVM, and reuse class-loading, linking, and method-profile data on the next run.
I built the full path with PostgreSQL and Podman: packaged training test, strict cache verification, a 20-run benchmark, and a target-specific container image. Local median readiness dropped from 1022.4 ms to 380.3 ms.
https://www.the-main-thread.com/p/quarkus-project-leyden-aot-cache
-
No Quarkus Insights today, but we'll be back next Monday, August 3rd, for a new session!
-
No Quarkus Insights today, but we'll be back next Monday, August 3rd, for a new session!
-
We're taking an additional holiday this coming Monday so no new Insights podcast but we'll be back August 3rd with a new session.
-
We're taking an additional holiday this coming Monday so no new Insights podcast but we'll be back August 3rd with a new session.
-
An upload bar reaching 100% proves that the bytes arrived. It says nothing about validation, importing, publication, or failure.
I rebuilt the pattern in Quarkus with two clocks: browser transfer progress, then a PostgreSQL-backed job processed by Scheduler and streamed over SSE. Reload-safe, cancellable, and honest about phases where no useful percentage exists.
https://www.the-main-thread.com/p/quarkus-long-running-job-progress-sse
-
An upload bar reaching 100% proves that the bytes arrived. It says nothing about validation, importing, publication, or failure.
I rebuilt the pattern in Quarkus with two clocks: browser transfer progress, then a PostgreSQL-backed job processed by Scheduler and streamed over SSE. Reload-safe, cancellable, and honest about phases where no useful percentage exists.
https://www.the-main-thread.com/p/quarkus-long-running-job-progress-sse
-
Vault's TOTP API can validate a six-digit code. That is not yet a complete MFA flow.
This Quarkus tutorial adds the missing boundaries: authenticated subject, stored enrollment mapping, strict input validation, a two-minute step-up JWT, normal RBAC, and a replay test against a patched Vault runtime.
https://www.the-main-thread.com/p/quarkus-vault-totp-step-up-auth
-
Vault's TOTP API can validate a six-digit code. That is not yet a complete MFA flow.
This Quarkus tutorial adds the missing boundaries: authenticated subject, stored enrollment mapping, strict input validation, a two-minute step-up JWT, normal RBAC, and a replay test against a patched Vault runtime.
https://www.the-main-thread.com/p/quarkus-vault-totp-step-up-auth
-
Sorry, we're on holiday so no Quarkus Insights today. Join us in two weeks for a new session!
-
Sorry, we're on holiday so no Quarkus Insights today. Join us in two weeks for a new session!
-
MCP tool authorization should not depend on the model following an instruction.
This builds a Quarkus MCP server where OPA evaluates skill manifests and caller context — trust tier, signature status, team, scopes — before a tool appears in tools/list or executes.
Rego compiles to Wasm and evaluates in-process. No sidecar, no network call. Each decision logs a policy version and reason codes.
-
MCP tool authorization should not depend on the model following an instruction.
This builds a Quarkus MCP server where OPA evaluates skill manifests and caller context — trust tier, signature status, team, scopes — before a tool appears in tools/list or executes.
Rego compiles to Wasm and evaluates in-process. No sidecar, no network call. Each decision logs a policy version and reason codes.
-
#AI can query your DB—but can it do it correctly? That’s the hard part. @MarcoBelladelli shows how #Hibernate + #Quarkus + #LangChain4j add validation & control back. Want fewer production surprises?
Dive in: https://javapro.io/2026/04/03/talk-to-your-data-natural-language-data-access-in-java/
-
#AI can query your DB—but can it do it correctly? That’s the hard part. @MarcoBelladelli shows how #Hibernate + #Quarkus + #LangChain4j add validation & control back. Want fewer production surprises?
Dive in: https://javapro.io/2026/04/03/talk-to-your-data-natural-language-data-access-in-java/
-
Java в эпоху ИИ. К чему пришли на пятом MTC Web Services Meetup для разработчиков
Еще несколько лет назад разработчики на Java в основном обсуждали новые инструменты для создания программ, особенности архитектуры приложений и способы сделать сервисы быстрее и надежнее. Сегодня к этим темам добавился еще один важный вопрос — как искусственный интеллект меняет профессию программиста. Именно с этого началась пятая встреча в рамках MTC Web Services Meetup, которая собрала в Москве Java-разработчиков, архитекторов и инженеров. Вместо привычной серии докладов организаторы открыли вечер большой дискуссией о том, как меняется современная разработка, чему теперь стоит учиться, какую роль играют ИИ-ассистенты и как их использование влияет на безопасность решений. Обо всем этом — в материале. Читать дальше
https://habr.com/ru/companies/ru_mts/articles/1059050/
#разработка #Retry_Policy #Spring_Data_JPA #производительность #Quarkus #митап #тренды_java #мтс_true_tech
-
⏰ Starting soon! Join us for Quarkus Insights Ep. #254 with Jeremy Davis for "Domain Driven Design and Hexagonal Architecture - Part 2"
http://bit.ly/quarkusinsights -
⏰ Starting soon! Join us for Quarkus Insights Ep. #254 with Jeremy Davis for "Domain Driven Design and Hexagonal Architecture - Part 2"
http://bit.ly/quarkusinsights -
Jib can put a Quarkus fast-jar on a small runtime image without a Dockerfile. That is the easy part.
The useful checks come afterward: Java 25, UID 65532, a direct Java entrypoint, no shell, stable application layers, and an amd64/arm64 manifest.
I worked through the full path with Red Hat Hummingbird and Podman:
-
Jib can put a Quarkus fast-jar on a small runtime image without a Dockerfile. That is the easy part.
The useful checks come afterward: Java 25, UID 65532, a direct Java entrypoint, no shell, stable application layers, and an amd64/arm64 manifest.
I worked through the full path with Red Hat Hummingbird and Podman:
-
510 daily posts on The Main Thread taught me two things: consistency works, and agents only help when verification keeps up.
I am changing the cadence now. Still Quarkus, Java, standards, and agentic engineering. More room for tests, examples, and claims that earned their confidence.
-
510 daily posts on The Main Thread taught me two things: consistency works, and agents only help when verification keeps up.
I am changing the cadence now. Still Quarkus, Java, standards, and agentic engineering. More room for tests, examples, and claims that earned their confidence.
-
Join us next Monday, July 13th, at 9am EST for Quarkus Insights Ep. 254 as Jeremy Davis returns to continue his discussion about building scalable monoliths with Quarkus, Domain-Driven Design, and Hexagonal Architecture.
-
Join us next Monday, July 13th, at 9am EST for Quarkus Insights Ep. 254 as Jeremy Davis returns to continue his discussion about building scalable monoliths with Quarkus, Domain-Driven Design, and Hexagonal Architecture.
-
Quarkus is a brutal repo for coding agents. Deployment, runtime, generated classes, and build items do not line up with grep.
I indexed the full platform tree with codebase-memory-mcp and ran the same @ConfigMapping trace in IBM Bob twice: grep versus graph MCP. Both got the answer. The graph path used fewer files and less context.
https://www.the-main-thread.com/p/quarkus-codebase-memory-mcp
-
Quarkus is a brutal repo for coding agents. Deployment, runtime, generated classes, and build items do not line up with grep.
I indexed the full platform tree with codebase-memory-mcp and ran the same @ConfigMapping trace in IBM Bob twice: grep versus graph MCP. Both got the answer. The graph path used fewer files and less context.
https://www.the-main-thread.com/p/quarkus-codebase-memory-mcp
-
What happens when one #AI call isn’t enough? You don’t add more prompts—you add agents. @kevindubois & Laura Cowen show how enterprise AI really scales.
Curious how production systems are built? Dive in: https://javapro.io/2026/03/31/agentic-ai-patterns-for-enterprise-software/
#Quarkus #LangChain4j #Microservices @QuarkusIO
-
What happens when one #AI call isn’t enough? You don’t add more prompts—you add agents. @kevindubois & Laura Cowen show how enterprise AI really scales.
Curious how production systems are built? Dive in: https://javapro.io/2026/03/31/agentic-ai-patterns-for-enterprise-software/
#Quarkus #LangChain4j #Microservices @QuarkusIO
-
🎉 New blog post: Binary file uploads to REST endpoints using #JakartaEE or #Quarkus.
How to build a #REST API that accepts binary attachments? #Java
Yet another blog post inspired by a question after one of my talks! So, folks, keep asking questions! 🙌
-
🎉 New blog post: Binary file uploads to REST endpoints using #JakartaEE or #Quarkus.
How to build a #REST API that accepts binary attachments? #Java
Yet another blog post inspired by a question after one of my talks! So, folks, keep asking questions! 🙌
-
Agent demos are easy to fake. The hard part is wiring the agent to systems you already have.
I validated Wanaku 0.1.3 locally: one MCP router, one imported HTTP toolset, one forwarded upstream MCP server, one merged catalog. The happy path works. A few local-mode features do not yet.
-
Agent demos are easy to fake. The hard part is wiring the agent to systems you already have.
I validated Wanaku 0.1.3 locally: one MCP router, one imported HTTP toolset, one forwarded upstream MCP server, one merged catalog. The happy path works. A few local-mode features do not yet.
-
You asked for it—it’s finally here! A full collection of #Java training videos: #SpringBoot #Quarkus #Helidon #Payara #EclipseStore #Microstream & ...
✔️ Build real understanding
✔️ Learn core concepts that still apply
✔️ Go beyond quick tutorialsWatch: https://youtube.com/playlist?list=PLFeSAZzYdUocK95uDSx7MPFi76GmkxPTm
-
You asked for it—it’s finally here! A full collection of #Java training videos: #SpringBoot #Quarkus #Helidon #Payara #EclipseStore #Microstream & ...
✔️ Build real understanding
✔️ Learn core concepts that still apply
✔️ Go beyond quick tutorialsWatch: https://youtube.com/playlist?list=PLFeSAZzYdUocK95uDSx7MPFi76GmkxPTm
-
A deep dive into time to peak performance with Leyden and Quarkus
And this is why, even if Java allows for developers to forget about memory management and low level coding (because the JVM itself is a master optimizer), that doesn't mean you should just blindly write code without knowing what happens behind the scenes.
Choosing the right dependencies can change your app performance.
https://quarkus.io/blog/when-the-jit-cant-keep-up/
#performance #optimization #OpenJDK #Leyden #AOTCache #Java #Quarkus #SpringBoot
-
A deep dive into time to peak performance with Leyden and Quarkus
And this is why, even if Java allows for developers to forget about memory management and low level coding (because the JVM itself is a master optimizer), that doesn't mean you should just blindly write code without knowing what happens behind the scenes.
Choosing the right dependencies can change your app performance.
https://quarkus.io/blog/when-the-jit-cant-keep-up/
#performance #optimization #OpenJDK #Leyden #AOTCache #Java #Quarkus #SpringBoot
-
Sorry, we're on holiday so no Quarkus Insights today. Join us next week for a new session!
-
Sorry, we're on holiday so no Quarkus Insights today. Join us next week for a new session!
-
Production incident and the first question is: have we seen this before?
I built a Quarkus service that turns Java incidents into vectors with deterministic feature hashing — no embedding model, no LLM. Store them in Qdrant, search by failure shape, filter by service and environment.
The vectors are inspectable and repeatable. The scores are explainable from the input.
New tutorial on The Main Thread:
-
Production incident and the first question is: have we seen this before?
I built a Quarkus service that turns Java incidents into vectors with deterministic feature hashing — no embedding model, no LLM. Store them in Qdrant, search by failure shape, filter by service and environment.
The vectors are inspectable and repeatable. The scores are explainable from the input.
New tutorial on The Main Thread:
-
The tenantId parameter in every service method is one of those patterns that works until it doesn't.
The Quarkiverse Multitenancy extension resolves the tenant once per request and gets out of the way. I built a small SaaS API with headers, JWT, and a custom host resolver to see how it holds up.
https://www.the-main-thread.com/p/quarkus-multi-tenant-extension
-
The tenantId parameter in every service method is one of those patterns that works until it doesn't.
The Quarkiverse Multitenancy extension resolves the tenant once per request and gets out of the way. I built a small SaaS API with headers, JWT, and a custom host resolver to see how it holds up.
https://www.the-main-thread.com/p/quarkus-multi-tenant-extension
-
We're starting our summer holiday bi-weekly schedule for Quarkus Insights so there won't be a session next week.
-
We're starting our summer holiday bi-weekly schedule for Quarkus Insights so there won't be a session next week.
-
SPDX gets more interesting when the SBOM is not only for scanners.
I wrote a hands-on Quarkus walkthrough around `quarkus-spdx`: generate an SPDX 3.0.1 SBOM, inspect the JSON-LD graph, follow declared license relationships, and check why NTIA/CISA minimums still fail out of the box.
https://www.the-main-thread.com/p/quarkus-spdx-sbom-json-ld-license-compliance
-
SPDX gets more interesting when the SBOM is not only for scanners.
I wrote a hands-on Quarkus walkthrough around `quarkus-spdx`: generate an SPDX 3.0.1 SBOM, inspect the JSON-LD graph, follow declared license relationships, and check why NTIA/CISA minimums still fail out of the box.
https://www.the-main-thread.com/p/quarkus-spdx-sbom-json-ld-license-compliance
-
Decision tables are good at deciding. They are terrible at waiting.
I wrote a Quarkus Flow 0.11 tutorial that keeps refund policy in plain Java and lets the workflow handle CloudEvents, correlation, manual-review callbacks, and persisted waiting state around the awkward "ask a human" path. https://www.the-main-thread.com/p/quarkus-flow-0-11-workflows-not-rules #Java #Quarkus #QuarkusFlow #CloudEvents #WorkflowEngines
-
Decision tables are good at deciding. They are terrible at waiting.
I wrote a Quarkus Flow 0.11 tutorial that keeps refund policy in plain Java and lets the workflow handle CloudEvents, correlation, manual-review callbacks, and persisted waiting state around the awkward "ask a human" path. https://www.the-main-thread.com/p/quarkus-flow-0-11-workflows-not-rules #Java #Quarkus #QuarkusFlow #CloudEvents #WorkflowEngines
-
Panache Next becoming Quarkus Data matters less than the rename and more than the API surface.
I wrote a hands-on tutorial that builds a small catalog API to show generated repositories, checked HQL, managed updates, stateless stock changes, paging, validation, and tests in one place. https://www.the-main-thread.com/p/quarkus-data-hibernate-java-persistence #Java #Quarkus #HibernateORM #JakartaData
-
Panache Next becoming Quarkus Data matters less than the rename and more than the API surface.
I wrote a hands-on tutorial that builds a small catalog API to show generated repositories, checked HQL, managed updates, stateless stock changes, paging, validation, and tests in one place. https://www.the-main-thread.com/p/quarkus-data-hibernate-java-persistence #Java #Quarkus #HibernateORM #JakartaData
-
A2A Java SDK 1.1.0.Final is live! Secure multi-tenant AI agents with new per-user task authorization & stability fixes: https://a2aproject.github.io/a2a-java/posts/a2a-java-sdk-1-1-0-final-released/ #Java #A2A #AIAgents #SoftwareEngineering #OpenSource #Quarkus #MultiTenancy
-
A2A Java SDK 1.1.0.Final is live! Secure multi-tenant AI agents with new per-user task authorization & stability fixes: https://a2aproject.github.io/a2a-java/posts/a2a-java-sdk-1-1-0-final-released/ #Java #A2A #AIAgents #SoftwareEngineering #OpenSource #Quarkus #MultiTenancy
-
⏰ Starting soon! Join us for Quarkus Insights Ep. #253: Some thoughts on Quarkus metrics with Bruno Baptista & Max Andersen.
-
⏰ Starting soon! Join us for Quarkus Insights Ep. #253: Some thoughts on Quarkus metrics with Bruno Baptista & Max Andersen.
-
“Delete this user” is where CRUD usually stops being CRUD.
This piece shows how Alex uses IBM Bob on a Quarkus app to map regression impact before implementation: active bookings, foreign keys, history reads, cache state, and which tests need to exist before the feature lands.