Praetorian Engineering Journal https://www.praetoriansolutions.uk/engineering/ Engineering investigations, decisions and operational lessons from real infrastructure work. en-gb Fri, 23 Oct 2026 21:40:00 GMT The benchmark was applied. The control model was not. https://www.praetoriansolutions.uk/engineering/notes/benchmark-applied-control-model-wasnt/ https://www.praetoriansolutions.uk/engineering/notes/benchmark-applied-control-model-wasnt/ Thu, 22 Oct 2026 00:00:00 GMT Hardening was commissioned as a governance exercise and delivered as a settings deployment. The disruption was visible; the missing access model was not. The backup succeeded. The restore had never been tested. https://www.praetoriansolutions.uk/engineering/notes/backup-succeeded-restore-never-tested/ https://www.praetoriansolutions.uk/engineering/notes/backup-succeeded-restore-never-tested/ Fri, 04 Sep 2026 00:00:00 GMT A green backup report confirmed that data had been copied. It did not prove that the service could be recovered. The dashboard was green because collection had stopped. https://www.praetoriansolutions.uk/engineering/notes/dashboard-green-collection-stopped/ https://www.praetoriansolutions.uk/engineering/notes/dashboard-green-collection-stopped/ Fri, 04 Sep 2026 00:00:00 GMT The last known state looked healthy. The monitoring system had quietly stopped learning anything new. Everything was monitored. Nobody owned the alert. https://www.praetoriansolutions.uk/engineering/notes/everything-monitored-nobody-owned-alert/ https://www.praetoriansolutions.uk/engineering/notes/everything-monitored-nobody-owned-alert/ Fri, 04 Sep 2026 00:00:00 GMT The condition was detected correctly. The operating model gave nobody a dependable reason to act on it. The failover worked. The service still failed. https://www.praetoriansolutions.uk/engineering/notes/failover-worked-service-still-failed/ https://www.praetoriansolutions.uk/engineering/notes/failover-worked-service-still-failed/ Fri, 04 Sep 2026 00:00:00 GMT The clustered platform moved exactly as designed. A dependency outside the cluster prevented the application from recovering. The firewall change was correct. The dependency map was not. https://www.praetoriansolutions.uk/engineering/notes/firewall-change-correct-dependency-map-wasnt/ https://www.praetoriansolutions.uk/engineering/notes/firewall-change-correct-dependency-map-wasnt/ Fri, 04 Sep 2026 00:00:00 GMT The rule matched the approved request. The request did not describe every service that depended on the traffic path. The working solution was replaced before its purpose was understood https://www.praetoriansolutions.uk/engineering/notes/working-solution-wasnt-understood/ https://www.praetoriansolutions.uk/engineering/notes/working-solution-wasnt-understood/ Thu, 03 Sep 2026 00:00:00 GMT The backups continued to run, but the recovery design, retention model and compliance position had been lost. The certificate did not expire unexpectedly https://www.praetoriansolutions.uk/engineering/notes/certificate-did-not-expire-unexpectedly/ https://www.praetoriansolutions.uk/engineering/notes/certificate-did-not-expire-unexpectedly/ Thu, 03 Sep 2026 00:00:00 GMT The expiry date was known years in advance. The outage happened because renewal ownership and deployment knowledge were not. The array was healthy. The disk was not. https://www.praetoriansolutions.uk/engineering/notes/array-was-healthy-disk-was-not/ https://www.praetoriansolutions.uk/engineering/notes/array-was-healthy-disk-was-not/ Thu, 03 Sep 2026 00:00:00 GMT The platform-level status remained reassuring while physical media telemetry showed that replacement planning should already have started. The missing agent exposed the missing baseline https://www.praetoriansolutions.uk/engineering/notes/missing-agent-exposed-missing-baseline/ https://www.praetoriansolutions.uk/engineering/notes/missing-agent-exposed-missing-baseline/ Thu, 03 Sep 2026 00:00:00 GMT A Proxmox telemetry warning revealed that a guest had been deployed outside the intended operational standard. The platform wasn't the decision https://www.praetoriansolutions.uk/engineering/notes/platform-wasnt-the-decision/ https://www.praetoriansolutions.uk/engineering/notes/platform-wasnt-the-decision/ Sat, 11 Jul 2026 00:00:00 GMT A vendor deadline created urgency. A working alternative was proven in the client environment within two hours. The contingency became the production platform https://www.praetoriansolutions.uk/engineering/notes/contingency-became-production/ https://www.praetoriansolutions.uk/engineering/notes/contingency-became-production/ Sat, 11 Jul 2026 00:00:00 GMT A temporary remote access path proved more dependable than the service it had been introduced to protect. The database wasn't the bottleneck https://www.praetoriansolutions.uk/engineering/notes/database-wasnt-the-bottleneck/ https://www.praetoriansolutions.uk/engineering/notes/database-wasnt-the-bottleneck/ Sat, 11 Jul 2026 00:00:00 GMT An inventory search fell from fifteen minutes to fifteen seconds when the application, SQL and infrastructure were treated as one platform. One source of truth mattered more than another system https://www.praetoriansolutions.uk/engineering/notes/single-source-of-truth/ https://www.praetoriansolutions.uk/engineering/notes/single-source-of-truth/ Sat, 11 Jul 2026 00:00:00 GMT A defence environment needed every operator to act from the same sonar tracking picture without conflicting information. The thirty-day project that took four https://www.praetoriansolutions.uk/engineering/notes/thirty-day-project-took-four/ https://www.praetoriansolutions.uk/engineering/notes/thirty-day-project-took-four/ Sat, 11 Jul 2026 00:00:00 GMT A network transformation estimated at thirty days was completed in four because the scope was understood before delivery began. The biggest vendor is still a single point of failure https://www.praetoriansolutions.uk/engineering/notes/biggest-vendor-still-single-point-of-failure/ https://www.praetoriansolutions.uk/engineering/notes/biggest-vendor-still-single-point-of-failure/ Sat, 11 Jul 2026 00:00:00 GMT CrowdStrike and Entra outages proved the same point: market leadership does not remove the need for independent recovery paths. The upgrade wasn't the objective https://www.praetoriansolutions.uk/engineering/notes/upgrade-wasnt-the-objective/ https://www.praetoriansolutions.uk/engineering/notes/upgrade-wasnt-the-objective/ Fri, 10 Jul 2026 00:00:00 GMT A platform upgrade preserved the client outcome while permanently changing the effort required to deliver it. The replacement wasn't the fix https://www.praetoriansolutions.uk/engineering/notes/replacement-wasnt-the-fix/ https://www.praetoriansolutions.uk/engineering/notes/replacement-wasnt-the-fix/ Thu, 09 Jul 2026 00:00:00 GMT A wireless refresh had been approved. The complaints survived the new access points. The cheapest tender became the most expensive decision https://www.praetoriansolutions.uk/engineering/notes/cheapest-tender-expensive-decision/ https://www.praetoriansolutions.uk/engineering/notes/cheapest-tender-expensive-decision/ Thu, 09 Jul 2026 00:00:00 GMT A discounted storage platform won the procurement. The architecture still had to work afterwards. Performance wasn't the problem https://www.praetoriansolutions.uk/engineering/notes/performance-wasnt-the-problem/ https://www.praetoriansolutions.uk/engineering/notes/performance-wasnt-the-problem/ Thu, 09 Jul 2026 00:00:00 GMT A switching refresh was proposed. Packet analysis showed the hardware was not the constraint. Technology isn't capability https://www.praetoriansolutions.uk/engineering/notes/technology-isnt-capability/ https://www.praetoriansolutions.uk/engineering/notes/technology-isnt-capability/ Wed, 08 Jul 2026 00:00:00 GMT Buying the platform is not the same as becoming able to operate it. The vendor has left. The work has not. https://www.praetoriansolutions.uk/engineering/notes/the-vendor-left/ https://www.praetoriansolutions.uk/engineering/notes/the-vendor-left/ Wed, 08 Jul 2026 00:00:00 GMT The right answer should still be the right answer when the vendor leaves the room.