A team has deployed a Lambda function that accidentally invokes itself, causing an unexpected spike in invocations and cost. According to AWS best practices, what should the team do immediately to stop the runaway invocations while they fix the code?