Page 3 - GitLab

Git

AI generated

Cache Flush vs. Cache Clean in Deployments: What You Actually Need and When

Maxim Mironjuk - October 20, 2025

Anyone who reflexively runs cache:flush after every deployment clears more than necessary and drives up load through needless cache regeneration. Precise cache invalidation after a release is a technical decision, not a habit.

Git

AI generated

Why You Should Never Build on Production: Artifacts, Reproducibility and Speed

Maxim Mironjuk - October 06, 2025

A build job on production is not a deployment, it is an uncontrolled experiment on the live system. Artifacts from isolated CI jobs bring reproducibility, speed and safety at the same time.

Git

AI generated

Documenting and Securing Manual Emergency Deployments

Maxim Mironjuk - October 02, 2025

When the pipeline is not fast enough, teams fall back on manual deployments. That is acceptable, but only if it produces complete documentation, a rollback path is prepared in advance, and the hotfix is cleanly folded back into the normal process afterward.

Git

AI generated

Artifact Signing, Checksums and Integrity for Releases

Maxim Mironjuk - October 02, 2025

A release artifact that has been altered without anyone noticing is more dangerous than a build that visibly fails. SHA256 checksums, GPG signatures and automated verification steps in the GitLab pipeline close this silent attack surface for Magento deployments.

Magento Git

AI generated

Zero Downtime in Magento: What Actually Works Without Outages and What Doesn't

Maxim Mironjuk - October 01, 2025

Zero downtime is promised often and defined rarely. This article explains honestly where a symlink switch truly works without any outage, and where database migrations unavoidably require a short maintenance window.

Magento Git

AI generated

Building Magento setup:upgrade Safely Into Your Deployment Pipeline

Maxim Mironjuk - September 30, 2025

setup:upgrade is the only command in the Magento deployment process that directly changes the production database. It does not run in the build job, it is not a test, it runs on the live system. A team that does not secure this step is building a deployment process that regularly needs luck when database migrations are involved.

Git

AI generated

Controlling Cron and Queue Consumers During Deployment

Maxim Mironjuk - September 26, 2025

Cron jobs and queue consumers still running during a deployment are an underestimated source of errors. A consumer that runs against the old release code while the symlink switches to the new release can cause data consistency problems. This article explains how to reliably control cron and consumers in the GitLab deployment process.

Git

AI generated

Redis, RabbitMQ, and OpenSearch in Zero-Downtime Deployments

Maxim Mironjuk - September 17, 2025

The symlink switch is not the end of a zero-downtime deployment: it is the start of a critical phase in which Redis, RabbitMQ, and OpenSearch must be compatible with the new code base. These three services are too often treated as peripheral details, even though they are frequently the root cause of post-deploy failures.

Git

AI generated

Runner Security: Shared Runner vs. Private Runner vs. Self-Hosted Runner

Maxim Mironjuk - September 13, 2025

Choosing the right runner type is not a matter of convenience, it is a security decision. Shared runners are practical, but their security model is often not sufficient for Magento production deployments involving SSH keys and database credentials. This article shows where the limits lie and what self-hosted runners can deliver.

Git

AI generated

Database Migrations for Zero Downtime: Expand/Contract, Compatibility, Risks

Maxim Mironjuk - September 07, 2025

No other part of a zero downtime deployment fails as often as database migrations. Renaming a column, dropping a table, adding a NOT NULL constraint: these are operations that are simply not compatible with code that is still running. This article explains the expand/contract pattern, when it works, and when a maintenance window is the more honest answer.