Salta al contenuto principale
Servizi

AEM audit and assessment: a technical review before you decide

An analysis of code, repository, indexes, workflows, Dispatcher and release process that ends with a prioritized backlog and a realistic roadmap.

An AEM audit exists to answer a simple question: what is the real state of our installation, and what should we do now? Intuitive answers are often wrong. An instance perceived as slow may suffer from a single query without an index, while an instance considered healthy may hide hundreds of unversioned OSGi configurations and /libs overlays that cannot be carried to the cloud. The AEM assessment replaces opinions with verifiable evidence.

We work on AEM Sites and AEM Assets, on any version and hosting model. The analysis is both static and dynamic: we read the code and the repository, but we also observe real behavior through logs, metrics, replication times and traffic profiles. When the context allows, we start from Best Practices Analyzer reports and verify them manually, because an automated report flags issues but does not decide priority.

The output is not a hundred-page document. It is a technical backlog in which each item has a description, evidence, risk, estimated effort and verification criterion, accompanied by a roadmap that separates immediate actions from those to be planned. Your team can take it on without us, or hand us the highest-impact items.

What an AEM audit analyzes

On the code side we examine the structure of the Maven modules (core, ui.apps, ui.config, ui.content, ui.frontend, dispatcher), the use of deprecated APIs, the quality of OSGi bundles, the presence of services with shared state, the correct use of service users instead of administrative sessions, and test coverage. We check Sling Models, HTL scripts and clientlibs to find business logic in the wrong place and duplication.

On the repository we analyze content structure, the presence of code and configuration under /content or /conf, /libs overlays, branch sizes, nodes with too many children, versions, and audit logs that grow unchecked. These symptoms determine both day-to-day performance and the cost of a future migration, and they are the ones that most often emerge only through direct inspection.

Oak indexes, queries and workflows

Slow queries are a frequent cause of degradation. We analyze query logs, check whether custom Oak indexes exist and whether the definitions are consistent with the Lucene indexing engine, and verify that the queries most used by components and servlets actually benefit from an index. On AEM as a Cloud Service indexes must be delivered as code, which makes this check even more important.

On workflows we verify which models run, how often, on how many payloads and with which custom steps. Overly heavy asset update workflows, steps that perform synchronous operations against external systems, and failed instances that were never cleaned up are typical problems. On Assets we also evaluate whether processing profiles and custom workflows are compatible with the cloud Asset Compute model.

Dispatcher, CDN and delivery

The Dispatcher configuration is often the least documented and most fragile part. We read farms, filters, cache rules, query parameter handling, rewrite rules and invalidation policies. We look for filters that are too permissive and expose paths not meant to be public, and rules that are too cautious and reduce the cache hit ratio. We verify how Cache-Control headers are set and how the CDN behaves in front of the Dispatcher.

We then measure the impact on user experience by observing LCP, INP and CLS on the main templates, distinguishing problems that originate on the AEM side (markup, clientlibs, images) from those caused by third-party tags. The detail of optimization work is described on the page about performance, Dispatcher and CDN.

Pipelines, environments and release process

A complete audit includes how code reaches production. On AEM as a Cloud Service we analyze Cloud Manager pipelines, the code quality, security and performance gates, environment variable management, and parity between development, stage and production. On on-premise installations we look at deployment procedures, packages, rollback strategies and release windows.

We also evaluate the editorial process: permission models, use of Launches, translation handling, rollout of Live Copies and Experience Fragments, and any use of Content Fragments. Often the cause of a technical problem is an editorial practice the system never governed, and ignoring it means the same problem will reappear after the cleanup.

What we deliver and what happens next

At the end we deliver a concise report, a backlog in a format that can be imported into the tool your team uses, a risk matrix and a roadmap with three horizons: low-effort immediate actions, items that can be planned over the coming quarters, and architectural decisions still to be made. A review session with architects and product owners closes the engagement, with room for questions and challenges.

After the assessment there are several paths. Some clients proceed internally, others hand us the code cleanup or the AEM as a Cloud Service migration, others activate managed services. In every case the backlog remains the client's property. The duration of an assessment depends on scope and should be estimated after an initial call.

FAQ

AEM audit FAQ

Normally read-only access to the code repository, logs, Dispatcher configuration and Cloud Manager, plus conversations with developers and administrators. For analysis of the live repository we use exported packages or non-production environments where possible.

No, it uses it as a starting point. The BPA flags patterns incompatible with AEM as a Cloud Service, but it does not set priorities or analyze Dispatcher, release process and editorial practices. Our work is to verify the results and turn them into an ordered backlog.

Yes. The same problems with indexes, workflows, caching and code affect on-premise installations too. The assessment helps reduce technical debt, prepare a version upgrade, or simply understand where maintenance cost is concentrated.

Of course. The assessment can be thematic: delivery only, Assets only, pipelines and code quality only. We define the scope together at the start so the result stays focused and consistent with the available budget.

The client. Backlog, risk matrix and roadmap are delivered in editable formats and can be used with or without our later involvement.

Request an AEM assessment

Tell us your version, hosting model and main problem. We will define an audit scope and tell you which access we need to start.

Talk to an AEM consultant