Github user HeartSaVioR commented on a diff in the pull request:
https://github.com/apache/storm/pull/2735#discussion_r197651740
--- Diff: storm-core/test/clj/org/apache/storm/metrics_test.clj ---
@@ -87,12 +91,22 @@
(first) ;; pick first task in the list, ignore other tasks' metric
data.
(second)
(or [])))
-
-(defmacro assert-buckets! [comp-id metric-name expected cluster]
- `(do
- (let [N# (count ~expected)]
- (wait-for-atleast-N-buckets! N# ~comp-id ~metric-name ~cluster)
- (is (= ~expected (subvec (lookup-bucket-by-comp-id-&-metric-name!
~comp-id ~metric-name) 0 N#))))))
+
+(defn assert-metric-running-sum! [comp-id metric-name expected min-buckets
cluster]
+ (try
+ (do
+ (wait-for-atleast-N-buckets! min-buckets comp-id metric-name cluster)
+ (.until
--- End diff --
So as far as I understand the description of issue, the metric value may
not be added to exactly N bucket though we simulate the time, so we need to
wait more to check if condition is met, and though we are now having another
waiting logic, we still have to call `wait-for-atleast-N-buckets!` because of
advancing simulated time, correct?
---
