DEVELOPER NEWS STREAM
Direct logs, engine updates, and framework notifications parsed from curated RSS feeds and announcements, updated hourly.

JH5世界上沒有串不起來的記憶體資源
圖片來源:Remote Direct Memory Access (RDMA)

James JoynerHow I Cut a Docker Image From 1.2GB to 180MB
A while back I inherited a service whose Docker image was 1.2GB. Pulls were slow, the CI cache was...

JH5如何用K8S Kueue & Cohort 有效的管理GPU/CPU資源 ?
在常見的AI 工作環境中,我們常需要在同一個 K8s 叢集上,同時運行給醫師使用的即時臨床推論 API,同時也要因應醫院內AI…

JH5如何用 Google OSS-Rebuild 打造一份符合 SaMD要求的 SBOM
如果是智慧醫療領域相關的人,這幾年應該很常聽到「軟體即醫療器材」(SaMD)的產品,你也一定對 SBOM (Software Bill of Materials,軟體物料清單) 這個詞不陌生。過去這可能只是個「建議事項」,但在美國 FDA…

JH5輕量級容器映像檔 CI/CD — Crane
過去的經驗,大部分工作與開發的環境大多具備了 crio / podman這些runtime,不過這兩年滿多平常手癢的專案其實都已經搬到 github codespace上,常常需要搬運映像檔、檢查 Manifest 內容,就會懶得裝環境,這時候,我都會想到 crane ,由…

JH5Docker/K8s 網路基礎:從手動操作 Linux Network Namespace 開始
在現代的雲端運算與容器化技術(如 Docker、Kubernetes)中,網路的隔離是一個基本的核心概念,在Multi-tenancy的環境中,我們不只希望隔離Process的檔案系統和Process ID,更重要的是要隔離彼此之間的網路環境,而Linux…

huangchengsirYou probably don't need ArgoCD - good-enough GitOps with git and docker compose
Every time someone starts self-hosting - a homelab, a few internal services for a small team - they...

TruffleDIRTY is yours to fix.
GitHub gives three reasons the merge button stays grey. Two of them are the maintainer's queue. One is yours, and the cost of letting it sit grows with the calendar.

Samson TanimawoBuilding Trust with Product Teams as an SRE
SRE teams that fight with product teams don't get things done. SRE teams that get along with product...

JH5如何用 Google OSS-Rebuild 打造一份符合 SaMD要求的 SBOM
如果是智慧醫療領域相關的人,這幾年應該很常聽到「軟體即醫療器材」(SaMD)的產品,你也一定對 SBOM (Software Bill of Materials,軟體物料清單) 這個詞不陌生。過去這可能只是個「建議事項」,但在美國 FDA…

Jakson TateMigrate Redis to DragonflyDB on Bare Metal: The Enterprise SRE Playbook
Security Notice: The Drop-In Replacement Myth Numerous promotional overviews state that...

Aasim GhaffarHow We Built an Atomic Circuit Breaker to Survive Third-Party API Outages at Scale
Every modern backend relies on third-party APIs for things like payment gateways, SMS notifications,...

ClickHouse system tables ate my disk (and the fix)
A monitoring alert said I was dropping check results. The disk was 100% full, but my real data was 20 MB. ClickHouse had logged 12 GB about itself. The config that caps it, and cuts idle CPU too.

JH5世界上沒有串不起來的記憶體資源
圖片來源:Remote Direct Memory Access (RDMA)

JH5輕量級容器映像檔 CI/CD — Crane
過去的經驗,大部分工作與開發的環境大多具備了 crio / podman這些runtime,不過這兩年滿多平常手癢的專案其實都已經搬到 github codespace上,常常需要搬運映像檔、檢查 Manifest 內容,就會懶得裝環境,這時候,我都會想到 crane ,由…

qingHow to Secure Your Linux Server in 10 Steps
How to Secure Your Linux Server in 10 Steps Introduction How to Secure Your...