You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Added include-retry-scenarios option to throughput_stress scenario (#229)
## What was changed
Added `include-retry-scenarios` option to `throughput_stress` scenario.
Setting `--option include-retry-scenarios=true` adds activities that
fail and retry, as actions to `throughput_stress`.
**NOTE**: enabling this option increases runtime of a single iteration
from ~23s to ~44s!
New activity actions were added to for this fail-and-retry coverage:
- retryable activity
- timeout activity
- heartbeat timeout activity
The added proto definitions are the same. I've separated them because
they cover different activity failure cases, which I think would be
neater to handle separately in the workers rather than a single failure
activity. It also allows them to be expanded separately and more
specifically.
I've added these actions to the existing throughput stress scenario, as
well as a cancellable activity action.
Added to corresponding logic to each worker to handle these new activity
actions.
## Why?
Provide additional activity coverage, particularly in cases where the
activity does not succeed on the first attempt.
2. How was this tested:
Ran the existing `throughput_stress_test.go` test.
Ran:
```
go run ./cmd run-scenario-with-worker \
--scenario throughput_stress \
--language go \
--iterations 10 \
--option internal-iterations=10 \
--option include-retry-scenarios=true
```
Having trouble running throughput stress on other language workers (even
before this change). Will address small fixes in subsequent PRs.
3. Any docs updates needed?
No
0 commit comments