Support for `resource.Retry` on mailgun API calls
tseho opened this issue · comments
short version:
Are you interested into receiving a PR with resource.Retry
on top of mailgun API calls, to improve robustness when their API have availability issues?
long version:
Hello,
Since a few days, mailgun is apparently under a DDOS attack.
https://status.mailgun.com/
While using this terraform provider for mailgun, we ended up with a few errors:
Error: while making http request: Get "https://api.mailgun.net/v3/domains/[redacted]/credentials?limit=100": dial tcp [redacted]:443: i/o timeout
or
Error: while making http request: Get "https://api.mailgun.net/v3/domains/[redacted]/credentials?limit=100": net/http: TLS handshake timeout
Even in DDOS conditions, we noticed that mailgun API was still available but with a higher error rate. (Which is understandable)
Nonetheless, encountering it at least once in automated workflows is an annoyance when a retry mechanism would have saved us from manually relaunching.
Would you be interested into having the terraform retry helper, at least on read operations like this one ?
If that's the case, I will try to find the time to open a PR with a proposal.
Thanks