<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Writing on Ilyas Hamdi</title><link>https://ilyashamdi.com/writing/</link><description>Recent content in Writing on Ilyas Hamdi</description><generator>Hugo -- gohugo.io</generator><language>en</language><lastBuildDate>Thu, 26 Mar 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://ilyashamdi.com/writing/index.xml" rel="self" type="application/rss+xml"/><item><title>The 100% reliability targets fallacy</title><link>https://ilyashamdi.com/writing/slo-100-percent-fallacy/</link><pubDate>Thu, 26 Mar 2026 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/slo-100-percent-fallacy/</guid><description>100% SLOs aren't just impractical. They're logically indefensible, and they corrode every reliability conversation that follows.</description></item><item><title>The AI build vs buy decision</title><link>https://ilyashamdi.com/writing/ai-build-vs-buy/</link><pubDate>Thu, 26 Feb 2026 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/ai-build-vs-buy/</guid><description>Build vs buy in AI is downstream of organizational position, not upstream of it. The decision answered without that context is mostly noise.</description></item><item><title>When AI makes decisions about learners</title><link>https://ilyashamdi.com/writing/ai-edtech-learners/</link><pubDate>Thu, 15 Jan 2026 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/ai-edtech-learners/</guid><description>Responsible AI in edtech needs the rigor of clinical or financial AI, plus a constraint those domains don't share: the population can't advocate for itself.</description></item><item><title>AI governance with no governance team</title><link>https://ilyashamdi.com/writing/ai-governance-without-team/</link><pubDate>Tue, 09 Dec 2025 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/ai-governance-without-team/</guid><description>The lightweight version of AI governance: four artifacts, named owners across functions you already have, before implicit governance is the only kind you've got.</description></item><item><title>Monitoring at scale</title><link>https://ilyashamdi.com/writing/monitoring-infrastructure-architecture-at-scale/</link><pubDate>Tue, 21 Oct 2025 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/monitoring-infrastructure-architecture-at-scale/</guid><description>The monitoring stack that worked at 50 services breaks at 500. Patterns that scale aren't bigger versions of the small ones. They're different patterns.</description></item><item><title>The error budget mindset</title><link>https://ilyashamdi.com/writing/error-budget-management-infrastructure/</link><pubDate>Tue, 22 Apr 2025 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/error-budget-management-infrastructure/</guid><description>The mechanism is the easy part. The willingness to slow shipping when the budget is exhausted is the hard part, and the part most programs skip.</description></item><item><title>Shaping architecture with SLOs</title><link>https://ilyashamdi.com/writing/slo-driven-infrastructure-architecture/</link><pubDate>Tue, 04 Mar 2025 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/slo-driven-infrastructure-architecture/</guid><description>SLO-driven architecture treats the SLO as upstream of the design. If your SLO has never caused you to ship later, it isn't doing the work.</description></item><item><title>The cost of telemetry</title><link>https://ilyashamdi.com/writing/infrastructure-telemetry-architecture/</link><pubDate>Tue, 14 Jan 2025 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/infrastructure-telemetry-architecture/</guid><description>Cardinality, retention, and query are the levers that decide what your team can know about the system, and how much knowing costs.</description></item><item><title>The implicit architecture of every modernization decision</title><link>https://ilyashamdi.com/writing/implicit-architecture-modernization/</link><pubDate>Tue, 19 Nov 2024 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/implicit-architecture-modernization/</guid><description>Every modernization arrives with a stated goal and a much larger implicit one. The modernization that fails is the one where the implicit architecture was never named.</description></item><item><title>The three stages every IDP goes through</title><link>https://ilyashamdi.com/writing/platform-engineering-internal-developer-platforms/</link><pubDate>Tue, 15 Oct 2024 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/platform-engineering-internal-developer-platforms/</guid><description>Most IDPs go through three painful stages: no users, wrong abstractions, and the moment someone realizes it has to be operated as a product.</description></item><item><title>Infrastructure security architecture patterns</title><link>https://ilyashamdi.com/writing/infrastructure-security-architecture-patterns/</link><pubDate>Mon, 19 Aug 2024 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/infrastructure-security-architecture-patterns/</guid><description>A security pattern adopted without a threat model is a control with no clear job. The patterns aren't the problem. The application is.</description></item><item><title>Zero trust network architecture</title><link>https://ilyashamdi.com/writing/zero-trust-network-architecture/</link><pubDate>Tue, 23 Jul 2024 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/zero-trust-network-architecture/</guid><description>Zero trust is one of the most successfully marketed architectural concepts of the decade. Most projects deliver better-than-perimeter, not actual zero trust.</description></item><item><title>What going containerized committed you to</title><link>https://ilyashamdi.com/writing/containerized-actual-commitment/</link><pubDate>Mon, 08 Apr 2024 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/containerized-actual-commitment/</guid><description>Containers were sold on consistency. The bill arrived in installments: orchestration, networking, security, observability. An honest audit of what the decision bought.</description></item><item><title>SLO targets and observability maturity</title><link>https://ilyashamdi.com/writing/slo-targets-observability-maturity/</link><pubDate>Mon, 18 Mar 2024 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/slo-targets-observability-maturity/</guid><description>SLO targets need to match observability maturity, not a generic standard. Sometimes leveling up observability is the precondition, not a parallel track.</description></item><item><title>The IDP that helps, and the one that just adds overhead</title><link>https://ilyashamdi.com/writing/idp-overhead/</link><pubDate>Tue, 12 Dec 2023 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/idp-overhead/</guid><description>The difference between a useful IDP and a costly one is rarely the technology. It's the discipline of building for actual users and admitting what's not working.</description></item><item><title>What a cloud architecture review should produce</title><link>https://ilyashamdi.com/writing/cloud-architecture-review-output/</link><pubDate>Thu, 26 Oct 2023 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/cloud-architecture-review-output/</guid><description>Most cloud architecture reviews produce a deck that gets filed. A useful review produces decisions framed for the people who can make them.</description></item><item><title>The security posture question most organizations ask too late</title><link>https://ilyashamdi.com/writing/security-posture-too-late/</link><pubDate>Mon, 14 Aug 2023 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/security-posture-too-late/</guid><description>Four triggers force the question 'what's our security posture, actually?' Three are expensive. The fourth is the one to engineer for yourself.</description></item><item><title>When compliance becomes a substitute for security thinking</title><link>https://ilyashamdi.com/writing/compliance-vs-security-substitute/</link><pubDate>Tue, 23 May 2023 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/compliance-vs-security-substitute/</guid><description>Compliance frameworks set a floor, not a ceiling. Treating them as the security strategy is how organizations end up audited and exploitable.</description></item><item><title>The postponed decision is now the risk</title><link>https://ilyashamdi.com/writing/postponed-technology-decision/</link><pubDate>Wed, 08 Mar 2023 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/postponed-technology-decision/</guid><description>The most expensive technology decisions are the postponed ones. They accumulate cost in places that don't show up as line items.</description></item><item><title>The architectural decisions hidden in a Dockerfile</title><link>https://ilyashamdi.com/writing/dockerfile-architectural-decisions/</link><pubDate>Tue, 21 Feb 2023 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/dockerfile-architectural-decisions/</guid><description>A Dockerfile is six lines and a dozen architectural commitments. Most teams write them by copy-paste. Few read them as decisions.</description></item><item><title>CI/CD pipeline architecture for infrastructure</title><link>https://ilyashamdi.com/writing/cicd-pipeline-architecture-infrastructure/</link><pubDate>Thu, 08 Dec 2022 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/cicd-pipeline-architecture-infrastructure/</guid><description>Application CI/CD is well-trodden. Infrastructure CI/CD is younger and stranger, and applying application patterns directly to it produces predictable failures.</description></item><item><title>IAM sprawl is access debt</title><link>https://ilyashamdi.com/writing/iam-sprawl-access-debt/</link><pubDate>Tue, 15 Nov 2022 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/iam-sprawl-access-debt/</guid><description>IAM is the part of your cloud footprint that grows fastest and gets cleaned up the slowest. The result looks like a control and behaves like a liability.</description></item><item><title>What technical leaders get wrong communicating architecture upward</title><link>https://ilyashamdi.com/writing/communicating-architecture-upward/</link><pubDate>Thu, 22 Sep 2022 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/communicating-architecture-upward/</guid><description>The same architecture decision can be approved by one audience and rejected by another in the same quarter. Which document each was reading is the difference.</description></item><item><title>What a cross-account migration forces you to confront</title><link>https://ilyashamdi.com/writing/cross-account-migration-confronts/</link><pubDate>Mon, 15 Aug 2022 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/cross-account-migration-confronts/</guid><description>A cross-account database move is the diagnostic, not the work. It surfaces architectural decisions you've been making implicitly for years.</description></item><item><title>IaC architecture at 5, 50, and 500 engineers</title><link>https://ilyashamdi.com/writing/iac-architecture-patterns/</link><pubDate>Tue, 14 Jun 2022 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/iac-architecture-patterns/</guid><description>The shape of your IaC architecture is downstream of team size. Most pain comes from running yesterday's architecture at today's scale.</description></item><item><title>The four kinds of cloud architecture debt</title><link>https://ilyashamdi.com/writing/cloud-architecture-debt/</link><pubDate>Tue, 19 Apr 2022 00:00:00 +0000</pubDate><guid>https://ilyashamdi.com/writing/cloud-architecture-debt/</guid><description>Cloud architecture debt has four shapes: governance, operational, knowledge, optionality. Read the shape before you decide what to pay down.</description></item></channel></rss>