feat: Automatic retry with exponential backoff

[Snider]

Summary

Automatic retry for transient failures with configurable backoff.

Use Case

Network hiccups, temporary 5xx errors, rate limit hits - should retry automatically.

Retry Strategy

Attempt 1: immediate
Attempt 2: wait 1s
Attempt 3: wait 2s
Attempt 4: wait 4s
Attempt 5: wait 8s (max)

Commands

# Default retry (3 attempts)
borg collect website https://example.com

# More attempts
borg collect website https://example.com --retries 5

# Custom backoff
borg collect website https://example.com --retry-backoff 2s --retry-max 30s

# No retry
borg collect website https://example.com --no-retry

Retryable Errors

Error	Retry?
429 Too Many Requests	Yes (with Retry-After)
500 Internal Server Error	Yes
502 Bad Gateway	Yes
503 Service Unavailable	Yes
504 Gateway Timeout	Yes
Connection timeout	Yes
Connection reset	Yes
DNS failure	Yes (1 retry)
404 Not Found	No
403 Forbidden	No

Configuration

# .borg-retry.yaml
retries: 5
backoff:
  initial: 1s
  multiplier: 2
  max: 60s
  jitter: 0.1  # ±10% randomization

# Per-error overrides
errors:
  429:
    use_retry_after: true
    max_wait: 300s
  dns:
    retries: 1

Options

Flag	Description
--retries N	Max retry attempts
--retry-backoff	Initial backoff duration
--retry-max	Maximum backoff duration
--retry-jitter	Randomization factor
--no-retry	Disable retries

Acceptance Criteria

• Exponential backoff implementation
• Retry-After header respect
• Jitter to prevent thundering herd
• Per-error-type configuration
• Max total retry time limit

[google-labs-jules]

Jules is on it. When finished, you will see another comment and be able to review a PR.

[google-labs-jules]

Ready for a review! A PR has been created.