Practice Exam

Explanations for each answer:

• Disable connection pooling and configure a second Azure Database for MySQL instance is incorrect. Disabling connection pooling would worsen the problem by forcing the application to create new connections for every request, increasing overhead and likelihood of connection failures. Connection pooling actually helps by reusing connections. Additionally, adding a second MySQL instance doesn't address transient connection errors—you need retry logic to handle temporary failures, not redundant infrastructure.
• Close the database connection and immediately report an error is incorrect. Immediately reporting errors without retry attempts defeats the purpose of transient fault handling. Transient errors are temporary conditions that often resolve within seconds. Best practices recommend waiting and retrying rather than immediately failing. Closing the connection is correct, but you should retry with a fresh connection after a delay, not report the error immediately.
• Wait five seconds before repeating the connection attempt to the database is correct. Waiting before retrying is a core component of transient fault handling. Microsoft recommends waiting at least 5 seconds before the first retry to avoid overwhelming the service during reconfiguration. This initial delay gives the database time to complete any ongoing maintenance or load balancing operations that caused the transient error.
• Increase connection repeat attempts exponentially up to 120 seconds is correct. Exponential backoff is the recommended retry strategy for Azure databases. After the initial 5-second wait, each subsequent retry should wait exponentially longer (e.g., 5s, 10s, 20s, 40s, 80s) up to a maximum of 60-120 seconds. This prevents overwhelming the database during longer outages while giving transient conditions time to resolve. The exponential increase reduces load while maximizing the chance of successful reconnection.
• Set a maximum number of connection attempts to 10 and report an error on subsequent connections is correct. Setting a maximum retry count prevents infinite retry loops and allows the application to fail gracefully when errors are not transient. After exhausting retries (typically 5-10 attempts), the application should report the error rather than continuing indefinitely. This ensures the application doesn't hang forever and can inform users or administrators about persistent connection problems.