Refactor GetBookmarkStatusFunction response logic

[babakamyljanovssw]

We are currently observing numerous 404 NotFound responses from the GetBookmarkStatusFunction in the client Application Insights. The current logic generates two types of responses:

• When bookmark is found:

if (bookmarks.Any())
{
    return req.CreateJsonResponse(new
    {
        error = false,
        message = "",
        bookmarkStatus = true,
    });
}

• When bookmark is not found:

return req.CreateJsonResponse(new
{
    error = false,
    message = "",
    bookmarkStatus = false,
}, HttpStatusCode.NotFound);

In both cases, the error property is set to false, which indicates a successful operation. The only distinction is the bookmarkStatus property and the different status codes (200 vs. 404). However, the front-end app relies on the bookmarkStatus property alone to display different icons, while Application Insights on the front end logs any 400-500 status codes as failures automatically.

Suggested Improvement

To streamline this logic, I’ve updated the response to a single format which makes it generic, returning a 200 OK status code in all cases. This change removes the 404 response for the "not found" case, as it conflicts with the error = false indicator and avoids marking these responses as failures.

var bookmarkStatus = bookmarks.Any();
if (bookmarkStatus)
{
    _logger.LogInformation($"Could not find results for rule id: {ruleGuid}, and user: {userId}");
}

return req.CreateJsonResponse(new
{
    error = false,
    message = "",
    bookmarkStatus,
});

Now, the response is a unified JSON object that differs only by the bookmarkStatus property, with a consistent 200 OK status code. This will ensure Application Insights does not log these responses as failures, aligning the backend response with frontend expectations.

Logging: The conditional _logger.LogInformation also logs when no bookmarks are found in back-end, making it easier to track missing bookmarks if needed.

Application Insights

Figure: The same function for different rules and users are logged as failures, in case it shouldn't be failure