Closed jakobht closed 1 month ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 68.14%. Comparing base (
2f08de5
) to head (73afab2
). Report is 5 commits behind head on master.:exclamation: Current head 73afab2 differs from pull request most recent head 557e21c
Please upload reports for the commit 557e21c to get more accurate results.
Files with Coverage Reduction | New Missed Lines | % | ||
---|---|---|---|---|
common/task/weighted_round_robin_task_scheduler.go | 1 | 88.06% | ||
common/task/parallel_task_processor.go | 2 | 93.06% | ||
service/matching/tasklist/db.go | 2 | 73.23% | ||
common/dynamicconfig/constants.go | 2 | 99.05% | ||
service/matching/tasklist/task_list_manager.go | 2 | 76.48% | ||
common/task/fifo_task_scheduler.go | 2 | 85.57% | ||
common/persistence/sql/sqlplugin/mysql/task.go | 2 | 73.68% | ||
common/membership/hashring.go | 2 | 84.69% | ||
common/persistence/sql/sqlplugin/mysql/db.go | 2 | 79.49% | ||
service/history/task/task.go | 3 | 84.81% | ||
<!-- | Total: | 299 | --> |
Totals | |
---|---|
Change from base Build 018fc614-f430-4db4-8939-669fc38230e0: | -0.1% |
Covered Lines: | 102545 |
Relevant Lines: | 147497 |
What changed? In the per workflow Id cache we have added request counting for each workflow Id for the external rate limits. We keep track of the number of requests for the current second for each workflow ID and emit this as a timer metric.
The timer metric keeps track of different percentiles of the metrics, as well as a max count, so we do not need to do this bookkeeping manually.
Why? We use the tracked information above to emit a metric stating the request count per workflow ID in a domain. We have to do the bookkeeping manually. Emitting a metric for each workflow ID would have way to high cardinality, and would overwhelm the metrics system.
How did you test it? Unit tests
Potential risks It should just emit a metric, so should be low risk.
Release notes
Documentation Changes