Commit 36af81b
committed
retries with tenacity for python sdk
Currently rety loop is done by hand and does not expose all required configuration args to the user i.e., for customizing health check backoffs, custom wait times, etc. This has resulted in issues where an error is raised even though it would have succeeded with a longer backoff. This PR introduces tenacity for more granualar retry control, introduces a standard exponential backoff pattern, and retains logic that captures spurious 502s as well as real 502s where the cluster is upgrading.
Signed-off-by: Anand Raman <anand.raman@feldera.com>
docs and timeout
tests and Retry-After header1 parent bdb2b29 commit 36af81b
10 files changed
Lines changed: 665 additions & 195 deletions
File tree
- python
- feldera
- rest
- tests/unit
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
1 | 1 | | |
| 2 | + | |
2 | 3 | | |
3 | 4 | | |
4 | 5 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
9 | 9 | | |
10 | 10 | | |
11 | 11 | | |
| 12 | + | |
0 commit comments