PinnedInDevelopersGlobalbyMoiz Ezzy·Jun 12Service Mesh Showdown: We Picked Istio. We’d Pick Linkerd Today.Three weeks. That’s how long we spent picking a service mesh before management called it. 40+ services on Kubernetes, 12 SREs on the team…
PinnedInDevelopersGlobalbyMoiz Ezzy·Jul 19 Linux Commands I’ve Run at 2AM More Times Than I Can CountIt was 2:07AM on a Tuesday in March when our payments API stopped accepting connections. Not slow. Not 5xx-ing. Refusing. The pod was up…
PinnedInDevelopersGlobalbyMoiz Ezzy·May 28Real SRE Interview Questions (and How I Answer Them)I have interviewed for SRE roles at startups and larger companies, and one pattern is consistent: interviewers are not testing…
InReaders ClubbyMoiz Ezzy·Nov 3I Read Google’s 84-Page SRE Draft. Here’s What’s Useful.The PDF of Google’s second SRE book that I received is 84 pages long. I opened its table of contents expecting twenty-two chapters of…
InTowards AWSbyMoiz Ezzy·Oct 24AWS Costs Keep Growing? Fix Them With a FinOps HabitWhen a team’s AWS bill jumps from $12,000 to $16,000 in a month, the first reaction is usually to hunt for one bad EC2 instance or an old…A response icon1A response icon1
InTowards AWSbyMoiz Ezzy·Oct 13Why AWS DevOps Agent Won’t Replace Your SRE TeamWhen PagerDuty fires at 2:13 AM, three dashboards disagree and the latest deployment looks suspicious.
InBeyond LocalhostbyMoiz Ezzy·Oct 9We Cut 60 GiB of Kubernetes Requests Without Breaking ProductionWe found 60 GiB of Kubernetes memory requests that our workloads rarely used.A response icon1A response icon1
InBeyond LocalhostbyMoiz Ezzy·Sep 10Our Kubernetes Incident Ran 4 Hours. Here Are the 7 Tools We Added.At 14:22 on a Thursday, our payment service started failing 8% of requests. It stayed at 8% for four hours while three of us stared at…
InBeyond LocalhostbyMoiz Ezzy·Sep 9One Digit Took Our Site Down. Now Chaos Runs on Every Deploy.Our circuit breaker was supposed to trip at 800ms. Someone typed 8000. That one extra zero took the site down three hours after the deploy…
InDevelopersGlobalbyMoiz Ezzy·Aug 21MTU Broke 3.2% of Our Requests for 6 Hours. Here’s the Trace.Our API gateway was failing exactly 3.2% of its calls to the auth service. Not spiking. Not recovering. 3.2%, flat as a ruler, for six…