Website maintenance checklist for stable growth

Record the domain, hosting, CMS or framework, databases, installed extensions and the external services the site depends on. Include email delivery, payment providers, scheduled jobs, analytics and storage. Identify the account owner and renewal responsibility for each service. This makes it easier to investigate a failure than starting with an unfamiliar login when something is already broken.

Rank the business journeys by the consequence of failure. A brochure site may depend mainly on accurate information and enquiries; a store also needs stock, payment and order handling. An internal application may need daily operational records. Agree maintenance responsibilities around those actual needs. The same calendar of plugin updates is not a complete care plan for every type of website.

Before a change, know what will be updated and how the previous state can be restored. Use staging when a release could affect custom functionality or commerce. Read relevant release information and pay particular attention to dependencies that interact with each other. Applying several unrelated changes at once can make an unexpected problem harder to diagnose.

Afterward, verify representative pages and the important workflows. Check that an editor can still publish, an enquiry reaches the correct destination and a supported purchase can complete in the appropriate test setup. Record what changed and what was checked. An update marked successful by the installer proves that files were replaced; it does not by itself prove that the site's business behaviour remained correct.

Decide what must be backed up and where the copy is stored. Files without the relevant database may not restore a dynamic website; a database without uploaded media may also be incomplete. Protect access to backups and clarify how long copies are retained. The business should understand who can restore the system if the usual developer or hosting contact is unavailable.

Perform a controlled restoration in a separate environment where practical. Verify that the restored application can access the required data and start correctly without affecting production. Note any external configuration, secrets or services needed during recovery. Choose backup frequency and acceptable recovery time according to the business impact of lost data and downtime, rather than relying on an impressive-sounding label without a tested procedure.

Availability checks can tell you that a page responds, but they may not catch a form that displays success without saving an enquiry. Review the outcomes that matter: accepted submissions, integration failures, scheduled tasks and renewal notices. Keep monitoring appropriate to the site's complexity and avoid collecting sensitive information merely because a logging tool makes it easy.

Define where an alert goes and who can act on it. An email notification has little value if nobody owns it or it arrives in an abandoned account. Separate a critical interruption from a routine content request so both can be handled sensibly. WordPress's monitoring handbook is a useful reference for that platform, but the operational question applies to any stack: how will a failure become visible before several customers report it?

Remove access that is no longer needed and review privileged accounts when staff or suppliers change. Check that the business still controls its domain, hosting and recovery options. Review page weight, mobile usability and third-party scripts after significant content or functionality changes. Routine care should notice gradual deterioration as well as sudden outages.

Keep new feature work distinct from maintenance. A useful backlog explains the problem, impact and priority of each improvement so the business can choose what to fund next. Agree how routine work is reported and what is outside the arrangement. This keeps expectations clear for a small site as well as a growing application and makes ongoing support a practical service rather than an unclear promise to handle everything.