Welcome to the thirty-fourth edition of Platform Engineering Monthly. As always, if you have suggestions or ideas for the next edition, please let me know!
📰 News
AWS says it can’t restore some data from mideast facilities struck by Iran
Seeing AWS say “can’t restore” in writing is pretty sobering. If you’ve been putting off multi-region for data you can’t afford to lose, this is your nudge.
Password spraying hit AWS root accounts at 150+ organizations
Another reminder to put MFA on your AWS root accounts and stop using them day to day. Password spraying root shouldn’t still work in 2026, and yet here we are.
Shai-Hulud now scans 469 credential locations, up from 189
Shai-Hulud’s been busy, going from 189 places it looks for credentials to 469. If you’ve still got long-lived tokens lying around on dev machines, now’s a good time to sort that out.
An undercover Google analyst infiltrated a notorious supply-chain hacking gang
Proper spy novel stuff. Worth a read if you want to see how these supply-chain gangs operate from the inside.
Kubernetes v1.37 can scale workloads to zero with the HorizontalPodAutoscaler
You can now scale to zero with the native HPA without bolting on extra tooling. It’s still beta in 1.37, but pretty nice for dev environments and anything that sits idle overnight burning money.
Claude, Codex, and Hermes installed unowned code inside corporate networks
Coding agents pulling in code nobody owns is basically shadow IT with extra steps. If you’re rolling these tools out, lock down what they’re allowed to install.
ChatGPT, Claude, and Grok all had outages at the same time
Three of the big AI providers went down within the same 90 minutes, each blamed a different unrelated fault and none of them has published a technical postmortem. If your platform leans on any of these, have a fallback.
Node 20 is no longer available in GitHub Actions
PSA if you’ve got actions pinned to older versions. The ACTIONS_ALLOW_USE_UNSECURE_NODE_VERSION escape hatch is gone, so anything still on Node 20 now breaks.
📚 Learning
20x the CI traffic without getting slower: how Datadog rebuilt Git serving
Great engineering write-up from Datadog. Git serving is one of those things nobody thinks about until CI traffic flattens it.
The Karpenter change that quietly turned on cross-AZ traffic charges
Many of us have been caught out by cross-AZ charges, and this one’s sneaky. Go check your Karpenter config and last month’s bill.
Why pod setup started timing out after 90 seconds
A proper look inside Cilium, with a real production problem at the end of it. These debugging write-ups are my favourite kind of post.
OpenTelemetry everywhere: Migrating a metrics platform at scale
Solid migration story. OTel’s pretty much the default now, so the hard bit is moving everything over without breaking the dashboards people rely on.
Kubernetes disaster recovery: guidance from three reproducible failure scenarios
Reproducible failure scenarios are the good stuff, a lot more useful than yet another DR doc nobody’s tested.
OpenTofu Is the Majority Terraform Engine on One Major Platform: What the 63% Tipping Point Means for Your Fleet’s IaC Layer
OpenTofu now handles 63% of runs and 72% of new workspaces on one major IaC platform. That’s one platform, not the whole market, but it’s a big number for a fork.
🧪 Interesting Projects
Cloud in a Bottle
Great name. I’m always keen on anything that lets you play with cloud stuff locally without a surprise bill at the end of the month.
Rootprint
Self-Hosted platform for storing OTEL logs. One to bookmark and poke at. I’ve not had a proper play with it yet, but it caught my eye.
restoredrill
Another reminder that you don’t know a backup works until you’ve restored it. Automating restore drills is the bit most of us skip.
cloudflare-github-actions-runner
GitHub Actions runners on Cloudflare is a fun idea. I wouldn’t bet prod on it yet, but I’m tempted to have a play.
PostgreSQL shared memory visualizer
Postgres shared memory is a bit of a black box for most of us, so a visualiser is pretty handy. Nice little tool.
Show HN: ZSvirt – A lightweight, scalable open source virtualization platform
Another open source virtualisation platform throwing its hat in the ring for the post-VMware crowd. Worth keeping an eye on whilst everyone’s shopping around.
📅 Events
KubeCon + CloudNativeCon North America 2026: Your Week in Salt Lake City
A handy logistics guide if you’re heading to Salt Lake City. Pack a water bottle, it’s a dry one.
DevOpsDays Philadelphia 2026
October 1-2, 2026 — Philadelphia, USA
Community-run DevOpsDays covering culture, automation and ops, plus open spaces.
Observability Summit Europe 2026
October 5, 2026 — Prague, Czechia
CNCF’s first Observability Summit in Europe. 25+ sessions across OpenTelemetry, Prometheus, Jaeger and Thanos.
DevOpsDays Prague 2026
October 5, 2026 — Prague, Czechia
One day, same week as Open Source Summit Europe, so handy if you’re already heading to Prague.
Devoxx Belgium 2026
October 5-9, 2026 — Antwerp, Belgium
Big developer conference with cloud, container and infrastructure tracks. Most relevant if your platform serves a lot of JVM teams.
Open Source Summit Europe 2026
October 7-9, 2026 — Prague, Czechia
The Linux Foundation’s main European event. Cloud infrastructure, Linux, security and a lot of the open source internal platforms are built on.
SREcon26 Europe/Middle East/Africa
October 13-15, 2026 — Dublin, Ireland
USENIX’s SRE conference for EMEA. Reliability, incident response, observability and running production at scale.
Kubernetes Community Days UK - Edinburgh 2026
October 19-20, 2026 — Edinburgh, UK
Two days of talks, project demos and workshops from teams running cloud native in production.
DevOpsDays Boston 2026
October 19-20, 2026 — Boston, USA
Two days on delivery, infrastructure automation and the organisational side of platform work.
Platform Engineering & Developer Experience Summit
October 21-22, 2026 — London, UK
Enterprise summit on internal developer platforms and developer experience, co-located with an Observability and SRE Summit.
Kubernetes Community Days Indonesia 2026
October 24, 2026 — Bandung, Indonesia
One day, 27+ sessions over four tracks on Kubernetes and cloud native.
GitHub Universe 2026
October 28-29, 2026 — San Francisco, USA
GitHub’s annual conference, in person and online. Expect a lot of AI-assisted development alongside Actions and developer tooling.
Have platform engineering tips to share? Reply to this email or connect with me on LinkedIn. Thanks for reading! Subscribe for free to receive new posts.