test: error handle, state mgmt, backoff, timeouts

[samrose]

What kind of change does this PR introduce?

EC2 Test Resilience Improvements

Retry Wrapper Function

• Added retry_with_backoff decorator that implements exponential backoff
• Configurable retry attempts, delays, and exception types
• Proper logging of retry attempts and failures

Error Handling and Logging

• Added comprehensive error handling throughout the code
• Improved logging with detailed messages and error context
• Added proper exception handling for AWS API calls

Instance State Management

• Added wait_for_instance_running function with retries
• Added proper state validation before proceeding
• Added timeout for instance state transitions

Backoff Strategy

• Implemented exponential backoff in the retry decorator
• Configurable initial delay and maximum delay
• Proper sleep intervals between retries

Resource Validation

• Added validate_aws_resources function to check security groups and IAM roles
• Validates resources before instance creation
• Provides clear error messages for validation failures

Simplified Startup

• Broke down the instance creation process into smaller, focused functions
• Each function has a single responsibility
• Better error isolation and handling

AWS API Timeouts

• Added proper timeouts for SSH connections
• Added timeout for health checks
• Added timeout for instance state transitions

Robust Health Checks

• Improved health check system with proper error handling
• Added timeout for health checks
• Better logging of health check failures
• Separate function for checking individual services

Cleanup Code

• Added proper cleanup in finally block
• Ensures instance termination even on failures
• Logs cleanup failures

Detailed Logging

• Added comprehensive logging throughout
• Logs all major operations and state transitions
• Logs errors with proper context
• Helps diagnose failures

[steve-chavez]

Will this help the sporadic timeouts we get from time to time on the testinfra CI job?

[samrose]

Will this help the sporadic timeouts we get from time to time on the testinfra CI job?

Steve, yes, I am trying to target that. I am going to wait on this until I finish https://github.com/supabase/postgres/pull/1547 as that will let me iterate on this locally