feat: implement API retry mechanism for Zoom data pipeline

[PiyushSonkusare]

What does this PR do?

• This PR improves the resilience of the backend data pipeline by introducing a reusable exponential backoff/retry utility for external API calls, specifically targeting the Zoom transcript ingestion scripts.

Specific Changes:

1. Utility Creation: Added pipeline/retryUtils.js to handle asynchronous retries globally. It defaults to 3 max retries with a 2000ms delay between attempts.

2. Pipeline Implementation: Wrapped the axios.post (OAuth token generation) and axios.get (Meeting summary fetch) inside pipeline/zoom-fetch-transcripts.js with the new withRetry wrapper.

Why is this needed?

• External services like the Zoom API occasionally drop connections, timeout, or rate-limit requests. Previously, a single dropped packet would cause the cron job to fail entirely for that execution. This retry mechanism catches those transient network errors, waits, and tries again, drastically reducing false alarms and the need for manual developer intervention.

How it was tested:

1. Created a local mock function to intentionally throw connection errors.

2. Verified that the wrapper successfully catches the error, waits the specified delay, and increments the retry count before ultimately failing on the final attempt.