Alerts reference
Overview
Alerts are the part of a CAST Profiler result that changes what you do next. Each one flags something about the source code delivery that will degrade a CAST Imaging analysis if it is left alone: code that is missing, code that will not be recognised, or code that is present more times than it should be. This page lists them all, with the reason each is raised and the action that resolves it.
Alerts are computed on every run unless you pass -na, which prevents the CAST Profiler CLI from calculating the alerts for which it is responsible. See the CLI reference.
Summary
| Alert | Raised when | Effect on a CAST Imaging analysis |
|---|---|---|
| Suspected missing SQL files | A persistence layer is detected but no .sql files are present |
Database structure and data access are absent from the results |
| No SQL file with CREATE TABLE | .sql files exist but none define schema |
No schema to anchor the data layer on |
| SQL in unrecognised file extensions | SQL code is found in files whose extension is not a known SQL extension | Those files are inventoried but not analyzed |
| Partially supported file extensions | Files carry an extension that is not standard for the language they contain | Those files are not analyzed |
| Unsupported data access technologies | Detected data access technologies have no analyzer coverage | The persistence and data access layers are incomplete |
| Duplicate files | The same file content appears at more than one path | Redundant objects make the results harder to read |
| Long paths | File paths exceed the supported length | Extraction can fail on those files |
| Generated files | Files carry a generated-code marker | Generated objects inflate or distort the results, depending on your goal |
| No JCL job identified | A mainframe application contains no JCL | Batch transaction call graphs cannot be built |
| High analysis complexity | The scored architecture complexity is high | Default configuration is unlikely to be optimal |
| Unsupported language volume | The share of the application in unsupported languages is above threshold | A material part of the application will not be analyzed |
Persistence and data access
Suspected missing SQL files
Why it is raised. A persistence layer was detected in the architecture but no .sql files were found. That combination means the delivery is probably incomplete rather than that the application has no database.
How to resolve it. Confirm that the .sql files containing the relevant database items were delivered. They may exist in a file format that needs converting, or be generated as part of a build script rather than committed. There is no need to deliver SQL files that contain only data.
No SQL file with CREATE TABLE
Why it is raised. SQL files are present, but none of them contain a CREATE TABLE statement. SQL script files sometimes carry data without any schema architecture - no tables, procedures or index structures - and the schema is what CAST Imaging needs to build the data layer and the architecture view.
How to resolve it. Deliver the SQL files that define the schema. As above, check whether they are held in a different format or produced by a build script.
The results also report a relational persistence score derived independently of the file extensions, based on CREATE TABLE statements and ORM entity mappings found anywhere in the code. Where that score indicates persistence but no schema files are present, the delivery is incomplete. Where it indicates no persistence at all, review the application design to confirm that no database is expected.
SQL in unrecognised file extensions
Why it is raised. SQL code was identified inside files whose extension is not one CAST Imaging treats as SQL. Those files will not be analyzed because the extension does not identify them as SQL.
How to resolve it. Rename the affected files to .sql before delivering the source code, and confirm the content is SQL. The affected files are listed in a downloadable report.
Unsupported data access technologies
Why it is raised. Data access or data access service technologies were detected that CAST Imaging does not support. This limits visibility into the persistence and data access layers.
How to resolve it. Review the listed technologies and check whether a supported alternative is in use elsewhere in the application, or contact CAST about extending coverage. The alert names the affected technologies so the conversation can be specific.
Source code delivery
Partially supported file extensions
Why it is raised. Files were found whose extension is not standard for the language they appear to contain. They will not be analyzed, because the extension is what selects the analyzer.
How to resolve it. Rename the affected files to one of the suggested extensions the alert lists, and confirm the file content matches the corresponding language.
Duplicate files
Why it is raised. The same file content was found at more than one path. Duplicated files complicate the visualization and interpretation of CAST Imaging results, and are frequently delivered by mistake - a vendored copy, a backup folder, or two branches exported side by side.
How to resolve it. Review the duplicates and remove the copies that were not intended to be part of the scope. The alert reports the number of files involved, the percentage of the delivery they represent, and the groups of paths that share content, downloadable as a report.
Duplicate detection compares content, not names, so it finds copies that were renamed. It runs across the whole scanned tree, including folders you may not have meant to deliver.
Long paths
Why it is raised. Files were found whose path exceeds the supported length. Long paths can be introduced by the extraction process itself, and can cause files to be skipped during a CAST Imaging analysis.
How to resolve it. Shorten the root path of the delivery, or contact your CAST administrator about the extraction configuration. The affected files are listed in a downloadable report.
Generated files
Why it is raised. Files carrying a generated-code marker were found. Whether they belong in the analysis depends entirely on what you are trying to achieve, so CAST Profiler reports them rather than deciding.
How to resolve it. Investigate the files and decide deliberately:
- For a modernization project, keep them. Generated code is part of what has to be understood and migrated.
- For an analysis focused on hand-written code and its quality, remove them from the scope, since generated objects will dominate the counts without telling you anything about the team’s code.
No JCL job identified
Why it is raised. No JCL was found in a mainframe delivery. Mainframe applications typically serve users through transactions and perform bulk processing in batch, and batch processing is driven by jobs written in JCL. CAST Imaging uses those jobs as the starting points for functional transactions involving programs and data.
How to resolve it. Confirm that all expected JCL files are part of the mainframe delivery. Without them, the batch transaction call graphs cannot be built, and a large part of what the application actually does stays invisible.
Scope and complexity
High analysis complexity
Why it is raised. The scored architecture complexity is high enough that the default analysis configuration is unlikely to give the best results.
How to resolve it. Continue with the analysis, but contact CAST for help with the configuration before you do. This alert is advice about tuning, not a blocker.
Unsupported language volume
Why it is raised. The share of the application written in languages with no CAST Imaging support is above the threshold CAST Profiler applies.
How to resolve it. Read the accompanying list of unsupported languages and frameworks, and take it into the scoping conversation with CAST. Depending on the languages involved, the options are an extension, a reduced scope, or accepting that one part of the application stays outside the knowledge base. The results page reports the supported percentage explicitly alongside its recommendation.
Reports
Findings can be exported from the results page as CAST Profiler findings Report.xlsx, which collects every alert in one workbook. Three alerts also have their own report, useful when the list is long enough to need filtering: duplicate files, long paths, and unknown SQL files.
In the derived results file, the same information is available under the alerts key - counts, boolean flags, and the arrays of affected paths - which is what to read if you are gating a pipeline on alerts rather than reading them by eye. See Result files.