Report exponential_avg_bucket_processing_time which gives more weight to recent buckets #43189

przemekwitek · 2019-06-13T08:54:09Z

Report exponential_avg_bucket_processing_time which gives more weight to recent buckets.

Related issue: #29857

elasticmachine · 2019-06-13T08:54:12Z

Pinging @elastic/ml-core

droberts195 · 2019-06-13T13:31:01Z

Now the timing_stats object is complete it would be good to add it to the docs in docs/reference/ml/apis/jobcounts.asciidoc.

Also, it would be good to add an example timing_stats (that's believable given the other stats) into the example response in docs/reference/ml/apis/get-job-stats.asciidoc.

benwtrent

I think adding docs and updating the yml REST tests would be good to do in this PR as well. My only comment in the review is not that important :D

benwtrent · 2019-06-13T20:32:43Z

.../src/main/java/org/elasticsearch/xpack/core/ml/job/process/autodetect/state/TimingStats.java

    /**
     * Updates the statistics (min, max, avg) for the given data point (bucket processing time).
     */
    public void updateStats(double bucketProcessingTimeMs) {
        if (bucketProcessingTimeMs < 0.0) {
-            throw new IllegalArgumentException("bucketProcessingTimeMs must be positive, was: " + bucketProcessingTimeMs);
+            throw new IllegalArgumentException("bucketProcessingTimeMs must be non-negative, was: " + bucketProcessingTimeMs);


#pedantry
So, is 0.0 considered a negative number? 😂

I'm not sure if you've seen the latest revision of this comment ("bucketProcessingTimeMs must be non-negative, was: ") or not. But now I believe it is correct i.e. if the value is less than 0 we throw an error that it should be non-negative (so 0 or greater than 0).

przemekwitek · 2019-06-14T10:04:42Z

Now the timing_stats object is complete it would be good to add it to the docs in docs/reference/ml/apis/jobcounts.asciidoc.

Done.

Also, it would be good to add an example timing_stats (that's believable given the other stats) into the example response in docs/reference/ml/apis/get-job-stats.asciidoc.

Done.

przemekwitek · 2019-06-14T10:05:04Z

I think adding docs and updating the yml REST tests would be good to do in this PR as well. My only comment in the review is not that important :D

Done.

droberts195 · 2019-06-14T10:16:50Z

docs/reference/ml/apis/jobcounts.asciidoc

+  (double) Maximum among all bucket processing times in milliseconds.
+
+`average_bucket_processing_time_ms`::
+  (double) Cumulative moving average of all bucket processing times in milliseconds.


Should this contain the word "moving"?

To match the wording of min and max it could just be "Average of all bucket processing times in milliseconds."

droberts195 · 2019-06-14T10:22:35Z

x-pack/plugin/src/test/resources/rest-api-spec/test/ml/jobs_get_stats.yml

+  - gt:     { jobs.0.timing_stats.minimum_bucket_processing_time_ms: 0.0 }
+  - gt:     { jobs.0.timing_stats.maximum_bucket_processing_time_ms: 0.0 }
+  - gt:     { jobs.0.timing_stats.average_bucket_processing_time_ms: 0.0 }
+  - gt:     { jobs.0.timing_stats.exponential_average_bucket_processing_time_ms: 0.0 }


It's probably safer to use gte on lines 104-107 and 143-146. The jobs are only creating results for 1 bucket, and if the bucket processing time for that 1 bucket happens to round down to 0 then the gt condition will fail.

droberts195

LGTM

przemekwitek · 2019-06-14T11:11:24Z

run elasticsearch-ci/default-distro

przemekwitek · 2019-06-14T11:55:22Z

run elasticsearch-ci/1

benwtrent

📊

… to recent buckets (elastic#43189)

…weight to recent buckets (#43189) (#43263)

przemekwitek added >enhancement WIP :ml Machine learning v8.0.0 v7.3.0 labels Jun 13, 2019

przemekwitek force-pushed the anomaly_detector_weighted_avg branch from 32552f0 to cdb756d Compare June 13, 2019 09:00

przemekwitek changed the title ~~Report timing stats as part of the Job stats response~~ Report exponential_avg_bucket_processing_time which gives more weight to recent buckets Jun 13, 2019

przemekwitek removed the WIP label Jun 13, 2019

przemekwitek marked this pull request as ready for review June 13, 2019 10:24

przemekwitek force-pushed the anomaly_detector_weighted_avg branch from 191295e to 3c91f8c Compare June 13, 2019 11:36

benwtrent reviewed Jun 13, 2019

View reviewed changes

droberts195 reviewed Jun 14, 2019

View reviewed changes

droberts195 approved these changes Jun 14, 2019

View reviewed changes

przemekwitek force-pushed the anomaly_detector_weighted_avg branch from fda922a to 945cc39 Compare June 14, 2019 11:27

benwtrent approved these changes Jun 14, 2019

View reviewed changes

przemekwitek added 6 commits June 14, 2019 21:02

Report timing stats as part of the Job stats response

39a56a6

Remove unused import

3197cf7

Add docs for timing_stats field

aa18e62

Update YML tests

1bab6da

Fix YML tests by allowing stats to be equal 0.0

8a9c508

Apply review comment

c322f1d

przemekwitek force-pushed the anomaly_detector_weighted_avg branch from 945cc39 to c322f1d Compare June 14, 2019 19:02

przemekwitek merged commit 13596c8 into elastic:master Jun 16, 2019

przemekwitek deleted the anomaly_detector_weighted_avg branch June 16, 2019 18:41

przemekwitek mentioned this pull request Jun 16, 2019

[7.x] Report exponential_avg_bucket_processing_time which gives more weight to recent buckets (#43189) #43263

Merged

przemekwitek added a commit to przemekwitek/elasticsearch that referenced this pull request Jun 16, 2019

Report exponential_avg_bucket_processing_time which gives more weight…

f459bcb

… to recent buckets (elastic#43189)

przemekwitek added a commit that referenced this pull request Jun 17, 2019

[7.x] Report exponential_avg_bucket_processing_time which gives more …

b2613a1

…weight to recent buckets (#43189) (#43263)

Mpdreamz mentioned this pull request Aug 7, 2019

[meta] 7.3 Release elastic/elasticsearch-net#4001

Closed

16 tasks

jakelandis added v8.0.0-alpha1 and removed v8.0.0 labels Jul 26, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Report exponential_avg_bucket_processing_time which gives more weight to recent buckets #43189

Report exponential_avg_bucket_processing_time which gives more weight to recent buckets #43189

przemekwitek commented Jun 13, 2019

elasticmachine commented Jun 13, 2019

droberts195 commented Jun 13, 2019

benwtrent left a comment

benwtrent Jun 13, 2019

przemekwitek Jun 14, 2019

przemekwitek commented Jun 14, 2019

przemekwitek commented Jun 14, 2019

droberts195 Jun 14, 2019 •

edited

Loading

przemekwitek Jun 14, 2019

droberts195 Jun 14, 2019

przemekwitek Jun 14, 2019

droberts195 left a comment

przemekwitek commented Jun 14, 2019

przemekwitek commented Jun 14, 2019

benwtrent left a comment

Report exponential_avg_bucket_processing_time which gives more weight to recent buckets #43189

Report exponential_avg_bucket_processing_time which gives more weight to recent buckets #43189

Conversation

przemekwitek commented Jun 13, 2019

elasticmachine commented Jun 13, 2019

droberts195 commented Jun 13, 2019

benwtrent left a comment

Choose a reason for hiding this comment

benwtrent Jun 13, 2019

Choose a reason for hiding this comment

przemekwitek Jun 14, 2019

Choose a reason for hiding this comment

przemekwitek commented Jun 14, 2019

przemekwitek commented Jun 14, 2019

droberts195 Jun 14, 2019 • edited Loading

Choose a reason for hiding this comment

przemekwitek Jun 14, 2019

Choose a reason for hiding this comment

droberts195 Jun 14, 2019

Choose a reason for hiding this comment

przemekwitek Jun 14, 2019

Choose a reason for hiding this comment

droberts195 left a comment

Choose a reason for hiding this comment

przemekwitek commented Jun 14, 2019

przemekwitek commented Jun 14, 2019

benwtrent left a comment

Choose a reason for hiding this comment

droberts195 Jun 14, 2019 •

edited

Loading