Certificate manager > Introduce Backoff Strategy for Certificate Creation and Renewal Retries

[hamadodene]

When a certificate creation or renewal fails, the system retries up to 10 times before marking the operation as permanently failed. However:

• There is no backoff logic in place.
• Retries occur on every manager cycle (every 5 minutes in our production setup).
• This results in all 10 retries being completed in about 50 minutes, which is too aggressive and inefficient (Too low for eventual remote issues).

Proposed Solution

Introduce a progressive backoff strategy for retries.

Backoff Intervals

5m, 5m, 10m, 15m, 30m, 1h, 2h, 4h, 8h, 8h

Behavior

• On failure:
  - Increment attempt_count.
  - Compute next_retry_at = now + backoff_interval[attempt_count].
• Manager will only retry if now >= next_retry_at.
• After 10 attempts, mark as FAILED.

[lblqgithub]

As you descriped,you already have a proposed solution,what could we do for you?