How packagecloud.io works
summary
packagecloud automatically extracts package metadata, atomically generates repository indexes and GPG‑signs each repo key held in AWS KMS, stores packages on S3 and serves them via Fastly; the stack uses Nginx, Rails/Sinatra, MySQL and Resque and favors vertical scaling on dedicated hardware. Practical wins include switching to TCMalloc for ~20% memory overhead reduction, setting TZ env var to avoid millions of syscalls, and spotting an EC2 syscall anomaly that makes time calls ~77% slower; overall the system works with multi‑layer redundancy but dedicated hosts increase maintenance risk.
AI-written summary · about
chapters (15)
- 0:00Intro
- 0:23Who am I
- 0:54Networking blog post
- 1:15What is package management
- 3:18What is the package repository
- 3:49What is packagecloud
- 5:07PackageSound Experience
- 6:13How packagecloud works
- 9:11Tips and tricks
- 10:25Vertical scaling tips
- 12:45System calls
- 14:37Monitoring
- 16:36Logging
- 18:59Primary Secondary DNS
- 19:43Dedicated Hardware
related talks
26mFuture is Cloud-native & Your Organization Should Be Too
15mSetup a Private, Secure, Cloud-native, Ultra-fast Repository in 60 sec w/Cloudsmith
29mCastle On A Cloud: The Github Story
37mTo Cloud or Not to Cloud?
21mAvoiding That $1M Dollar AWS Bill - Mark Michael, Glenn Oppegard
40mDeploying any Rails application to any cloud in minutes
21mHybrid Cloud - nothing to worry about! 🌥️
38mBuilding a Cloud Native Culture in an Enterprise - Deep Kapadia & Tony Li