Skip to content

Conversation

@lcawl
Copy link
Contributor

@lcawl lcawl commented Aug 13, 2018

This PR adds items to the 6.4.0 changelog that were identified by the https://github.com/elastic/ml-cpp/blob/master/dev-tools/mlcpp-release-notes.pl tool. It also contains some edits to the existing items.

When this list is satisfactory, it will be added to the Elasticsearch Release Notes.

NOTE: None of these items appear in the 6.4.0 Release Highlights (https://www.elastic.co/guide/en/elasticsearch/reference/6.4/release-highlights-6.4.0.html) yet. I assume the rules feature needs to be added there, but if any others also need to be highlighted, let me know!

@lcawl lcawl added the :ml label Aug 13, 2018
Copy link

@droberts195 droberts195 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks fine except for notes that people put in the wrong section.

* Corrects the model bounds and typical value calculation for time series models
which use a multimodal distribution. This issue could cause "Unable to bracket
left percentile =..." errors to appear in the logs. ({ml-pull}178[#178], {ml-pull}176[#176])
* Fixes a SIGSEGV in the autodetect process when jump upgrading from 5.6 to 6.3 ({ml-pull}144[#144], {ml-pull}143[#143])

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This one is already released and release-noted in 6.3.2.

left percentile =..." errors to appear in the logs. ({ml-pull}178[#178], {ml-pull}176[#176])
* Fixes a SIGSEGV in the autodetect process when jump upgrading from 5.6 to 6.3 ({ml-pull}144[#144], {ml-pull}143[#143])
* Fixes issues upgrading the state, which could cause the autodetect process to crash.
({ml-pull}139[#139], {ml-issue}136[#136](issue: {ml-issue}135[#135]))

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This one is already released and release-noted in 6.3.2.

* Fixes a SIGSEGV in the autodetect process when jump upgrading from 5.6 to 6.3 ({ml-pull}144[#144], {ml-pull}143[#143])
* Fixes issues upgrading the state, which could cause the autodetect process to crash.
({ml-pull}139[#139], {ml-issue}136[#136](issue: {ml-issue}135[#135]))
* Fixes by fields such that they respect `model_plot_config.terms` ({ml-pull}86[#86] (issue: {ml-issue}30004[#30004]))

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This one is already released and release-noted in 6.3.0.

({ml-pull}139[#139], {ml-issue}136[#136](issue: {ml-issue}135[#135]))
* Fixes by fields such that they respect `model_plot_config.terms` ({ml-pull}86[#86] (issue: {ml-issue}30004[#30004]))
* Function description for population lat_long results should be lat_long
instead of mean ({ml-pull}81[#81] (issue: {ml-issue}80[#80]))

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This one is already released and release-noted in 6.3.0.

* Checks accesses are in bounds when clearing recycled models
({ml-pull}79[#79] (issue: {ml-issue}76[#76]))
* Fixes error causing overestimation of effective history length
({ml-pull}66[#66] (issue: {ml-issue}57[#57]))

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This one is already released and release-noted in 6.3.0.

* Function description for population lat_long results should be lat_long
instead of mean ({ml-pull}81[#81] (issue: {ml-issue}80[#80]))
* Checks accesses are in bounds when clearing recycled models
({ml-pull}79[#79] (issue: {ml-issue}76[#76]))

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This one is already released and release-noted in 6.3.0.

Fix corner case failing to calculate lgamma values and the correspoinding log errors ({pull}#126[#126])
Influence count per bucket for metric population analyses was wrong and lead to wrong influencer scoring ({pull}#150[#150])
Fix a possible SIGSEGV for jobs with multivariate by fields enabled which would lead to the job failing ({pull}#170[#170])
* Detectors now support {stack-ov}/ml-rules.html[custom rules] that enable the user to improve machine learning results by providing some domain-specific knowledge in the form of rule. ({ml-pull}119[#119])
Copy link

@droberts195 droberts195 Aug 14, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This change will also probably be picked up by the release note script on the Java side - the two notes should be merged when the C++ changelog is merged with the Java one.

@lcawl
Copy link
Contributor Author

lcawl commented Aug 14, 2018

I just checked with @pcsanwald and they're not going to repeat backported PRs in the Elasticsearch Release Notes anymore, so I'll remove those duplicates. Thanks!

Copy link

@droberts195 droberts195 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@lcawl lcawl merged commit 23ab831 into elastic:6.4 Aug 14, 2018
@lcawl lcawl deleted the lcawley-64-RN branch August 14, 2018 17:20
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants