Extend EMR System Test waiter timeout#31071
Conversation
| task_id="add_steps", | ||
| job_flow_id=create_job_flow.output, | ||
| steps=SPARK_STEPS, | ||
| wait_for_completion=True, |
There was a problem hiding this comment.
I just moved this outside of the example code block (to line 170 below) as done elsewhere. We've been trying to clean up the snippets that get pasted into the docs but must have missed this one. No functionality change
| add_steps.waiter_delay = 60 | ||
| add_steps.waiter_max_attempts = 90 |
There was a problem hiding this comment.
Default values are 30 and 60
There was a problem hiding this comment.
The new config means we'll wait an hour and half, which feels a bit too long? That would block the whole test suite from completing for that long if this step hung.
If we used to wait 30min total I'd bump it up to 40min or 45min, anything beyond that we really need to find some other fix I think 😬
There was a problem hiding this comment.
Sure, sounds good. I'll dial these back a bit. I feel like maybe adding more retries at the default duration then? That would mean we can still succeed sooner, but get to try longer.
On rare occasion this system test has timed out. Extending the max wait to try to mitigate that.
cc: @o-nikolas @vincbeck @vandonr-amz @syedahsn