Introduce new setting to configure when to build graph during segment creation #2007

VijayanB · 2024-08-27T18:45:26Z

Description

Added new updatable index setting "build_vector_data_structure_threshold", which will be considered when to build relevant vector data structure or not for native engines before creating and merging segments. The default value is chosen to be 0 to make it backward compatible. This is applicable only for native engines, and, it is no-op for lucene.
We also don't need a feature flag since implementation is behind NativeEngines990KnnVectorsFormat. Hence, only index that are created 2.17 and above will have the ability to pause building graph while indexing or merging based on this updatable index setting.

In this PR, we introduced a setting to decide when to build the graph with min as -1 ( no graph will be built ) and Integer.MAX_Value - 2 ( as max value up to which user can provide ). Currently the default value is 0 ( build graph always ). We will be updating this value to a reasonable value ( greater than 0) to optimize when should we build graph if user didn't provide this value as next PR.

Similarly, during search, if no graph is available, we will call exact search for conssitency.

Related Issues

Part of #1942

Check List

New functionality includes testing.
New functionality has been documented.
API changes companion pull request created.
Commits are signed per the DCO using --signoff.
Public documentation issue/PR created.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

navneet1v · 2024-08-27T20:18:54Z

src/main/java/org/opensearch/knn/index/KNNSettings.java

@@ -65,6 +65,7 @@ public class KNNSettings {
     * Settings name
     */
    public static final String KNN_SPACE_TYPE = "index.knn.space_type";
+    public static final String KNN_BUILD_GRAPH = "index.knn.build_graph";


Suggested change

public static final String KNN_BUILD_GRAPH = "index.knn.build_graph";

public static final String KNN_BUILD_GRAPH = "index.knn.build_vector_datastructure";

navneet1v · 2024-08-27T20:19:15Z

src/main/java/org/opensearch/knn/index/KNNSettings.java

@@ -97,6 +98,7 @@ public class KNNSettings {
     */
    public static final boolean KNN_DEFAULT_FAISS_AVX2_DISABLED_VALUE = false;
    public static final String INDEX_KNN_DEFAULT_SPACE_TYPE = "l2";
+    public static final boolean INDEX_KNN_DEFAULT_BUILD_GRAPH = true;


rename this as per the new setting name I suggested on top.

navneet1v · 2024-08-27T20:21:14Z

src/main/java/org/opensearch/knn/index/codec/BasePerFieldKnnVectorsFormat.java

+    private boolean getBuildGraphSetting() {
+        // Default should be true, only explicit false should disable building graph for
+        // native engine
+        if (mapperService.isEmpty()) {
+            return true;
+        }


if MapperService is empty the code should have crashed in line 77. I think we should get the mapperservice from line 77 and pass it in the function.

navneet1v · 2024-08-27T20:21:50Z

src/main/java/org/opensearch/knn/index/codec/BasePerFieldKnnVectorsFormat.java

+            return true;
+        }
+        IndexSettings indexSettings = mapperService.get().getIndexSettings();
+        return indexSettings.getValue(KNNSettings.INDEX_KNN_BUILD_GRAPH_SETTING);


we should pass a default value if the setting is not found.

only index that are created 2.17 and above will have the ability to pause building graph

Do we need a index version check in this? If an index is created in 2.16 and then when the cluster is upgraded this will still allow the feature right?

navneet1v · 2024-08-27T20:22:39Z

src/main/java/org/opensearch/knn/index/codec/KNN990Codec/NativeEngines990KnnVectorsFormat.java

+        this(lucene99FlatVectorsFormat, true);
+    }
+
+    public NativeEngines990KnnVectorsFormat(final FlatVectorsFormat lucene99FlatVectorsFormat, boolean buildGraphSetting) {


Suggested change

public NativeEngines990KnnVectorsFormat(final FlatVectorsFormat lucene99FlatVectorsFormat, boolean buildGraphSetting) {

public NativeEngines990KnnVectorsFormat(final FlatVectorsFormat lucene99FlatVectorsFormat, boolean buildVectorDataStructure) {

navneet1v · 2024-08-27T20:23:42Z

src/main/java/org/opensearch/knn/index/codec/KNN990Codec/NativeEngines990KnnVectorsWriter.java

+        if (!buildGraph) {
+            return;


we should have an info log here. Other wise we will never know if the vector data structures are created or not.

navneet1v · 2024-08-27T20:23:48Z

src/main/java/org/opensearch/knn/index/codec/KNN990Codec/NativeEngines990KnnVectorsWriter.java

+        if (!buildGraph) {
+            return;


same as above.

navneet1v · 2024-08-27T20:24:48Z

src/testFixtures/java/org/opensearch/knn/KNNRestTestCase.java

+    protected Settings buildKNNSettingsWithBuildGraphProperty(boolean buildGraph) {
+        return Settings.builder()
+            .put("number_of_shards", 1)
+            .put("number_of_replicas", 1)


why 1 replica? this will keep the cluster yellow. If you want to test with more than 1 shard do that.

jmazanec15 · 2024-08-27T20:34:52Z

@VijayanB What is behavior around search before, mixed and after this setting is set from false to true? Are some segments search via exact and others not? Also, why is this an index setting and not field mapping parameter?

shatejas

@VijayanB The primary advantage seems to be reduce the CPU spike before merge, Are we seeing any improvements during force merge? Can you add the CPU usage of before and after with this PR? we probably need one for index and one for force merge

Overall can we have a summary of alternative solutions considered? I am a bit hesitant to add a user exposed setting which does not give the improvements out of the box.

shatejas · 2024-08-27T20:37:05Z

src/main/java/org/opensearch/knn/index/codec/BasePerFieldKnnVectorsFormat.java

+            return true;
+        }
+        IndexSettings indexSettings = mapperService.get().getIndexSettings();
+        return indexSettings.getValue(KNNSettings.INDEX_KNN_BUILD_GRAPH_SETTING);


only index that are created 2.17 and above will have the ability to pause building graph

Do we need a index version check in this? If an index is created in 2.16 and then when the cluster is upgraded this will still allow the feature right?

heemin32 · 2024-08-27T21:57:34Z

I could not make a connection between the issue #1942 and this PR. How is this PR going to be used for the issue?
About delaying of graph creation, shouldn't it be more automated instead user manually turn it on and off?

navneet1v · 2024-08-28T00:55:17Z

I could not make a connection between the issue #1942 and this PR. How is this PR going to be used for the issue? About delaying of graph creation, shouldn't it be more automated instead user manually turn it on and off?

@heemin32 with this change the solution builds the capability to completely stop the Vector DataStructure creation, which is Delaying Vector Data structures Creation in the GH issue.

Now the GH issue says that graph creation should be done even after force_merge, but as of now it cannot be done, because it requires coming with new API for doing merges. Plus the way force_merge works with 1 segment use-case is it doesn't build temporary segments and directly merges to 1 segment.

So with this capability the idea would be if we are doing indexing once, we ingest data by disabling graph creation and then enable the setting. After that we do force_merge to 1 segment for best performance for index build.

The next part of the issue will be to have a vector threshold based mechanism to see when to build the graph and when not to.

To avoid confusion I added more details on the GH issue.

navneet1v · 2024-08-28T00:56:03Z

src/main/java/org/opensearch/knn/index/KNNSettings.java

+     * build_vector_datastructure - This parameter determines whether to build vector data structure for knn fields during indexing
+     * and merging.
+     */
+    public static final Setting<Boolean> INDEX_KNN_BUILD_VECTOR_DATASTRUCTURE = Setting.boolSetting(


should we build this setting as int rather than bool, so that when we have threshold based graph build same setting can be used.

we can min value as -1 and max 1 to start with. and in next time we can increase the max value to be very high

navneet1v · 2024-08-28T01:04:47Z

src/main/java/org/opensearch/knn/index/codec/KNN990Codec/NativeEngines990KnnVectorsWriter.java

+        if (!buildVectorDatastructure) {
+            log.info("Skip building vector datastructure");
+            return;
+        }


we should add a codec level tests which will ensure that if this setting is true in reality graphs are not created.

VijayanB · 2024-08-28T04:33:22Z

I could not make a connection between the issue #1942 and this PR. How is this PR going to be used for the issue? About delaying of graph creation, shouldn't it be more automated instead user manually turn it on and off?

As a part of incremental graph build approach, this PR will allow user to have ability to pause building graph before completing ingestion process. Before this feature, knn plugin will build graph every time during flush, which contributes to the high total indexing time whenever new segments are being created. If user is going to merge segments after ingestion, those graphs which are created during segment flush will be discarded and new graph will be built from scratch during force merge. This gives additional control to users who wishes to reduce indexing time if their use case doesn't require building graph on every segment creation step.

As follow up, we will provide automated solution to decide whether to build graph or not based on number of vectors in given segment. We will introduce a threshold parameter and will let flush/merge decide whether to build graph or not accordingly.

VijayanB · 2024-08-28T05:13:20Z

@VijayanB The primary advantage seems to be reduce the CPU spike before merge, Are we seeing any improvements during force merge? Can you add the CPU usage of before and after with this PR? we probably need one for index and one for force merge

Overall can we have a summary of alternative solutions considered? I am a bit hesitant to add a user exposed setting which does not give the improvements out of the box.

The primary advantage is to provide more control to users on when to build vector data structures like graph. In current approach, CPU utilization for indexing is high due to graph creation process which will be later discarded if force merge is executed after ingestion. This setting will allow users to disable graph building step while ingestion ( this will reduce CPU utilization, which in turn improve ingestion throughput), and later update setting before merge in order to build graph as part of merge step, just before search operation.

VijayanB · 2024-09-04T20:59:10Z

Updated this PR's base to feature branch

Previosuly we were not creating hnsw file on segment flush for faiss engine. After successfully integrating hnsw file creation, we forgot to update unit test. Here, we will confirm that required files are being created based on field type. Signed-off-by: Vijayan Balasubramanian <[email protected]>

Added new updatable index setting "build_vector_data_structure_threshold", which will be considered when to build braph or not for native engines. This is noop for lucene. This depends on use lucene format as prerequisite. We don't need to add flag since it is only enable if lucene format is already enabled. Signed-off-by: Vijayan Balasubramanian <[email protected]>

shatejas · 2024-09-16T18:40:07Z

src/main/java/org/opensearch/knn/index/codec/KNN990Codec/NativeEngines990KnnVectorsWriter.java

-                KNNVectorValues<?> knnVectorValues = getVectorValues(vectorDataType, field.getDocsWithField(), field.getVectors());
+            if (totalLiveDocs == 0) {
+                log.debug("[Flush] No live docs for field {}", fieldInfo.getName());
+                return;


The return will break multi-field case since this is inside a for loop, I thought it was trivial enough to not add a test case, I was probably wrong. Maybe add List.of(Map.of(//some vectors), Map.emptyMap(), Map.of(//some vectors)) in the test case so we cover it (before/after) empty case in 1 case.

Same for line 107

Sure. I will update the test case

shatejas · 2024-09-16T18:46:28Z

...a/org/opensearch/knn/index/codec/KNN990Codec/NativeEngines990KnnVectorsWriterFlushTests.java

+        }
+    }
+
+    public void testFlush_whenThresholdIsEqualToNumberOfVectors_thenNativeIndexWriterIsCalled() throws IOException {


Should we add a case where vectorSize > threshold and verify only threshold vectors are added?

Added unit test

...a/org/opensearch/knn/index/codec/KNN990Codec/NativeEngines990KnnVectorsWriterFlushTests.java

Signed-off-by: Vijayan Balasubramanian <[email protected]>

shatejas

LGTM!

vibrantvarun · 2024-09-17T07:16:18Z

CHANGELOG.md

@@ -6,6 +6,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),

 ## [Unreleased 3.0](https://github.com/opensearch-project/k-NN/compare/2.x...HEAD)
 ### Features
+* Introduce new setting to configure whether to build vector data structure or not during segment creation [#2007](https://github.com/opensearch-project/k-NN/pull/2007)


Why this change log is added under unreleased 3.0?

Good catch. I will add it to 2.x when i create PR to 2.x Currently this PR is raised against feature branch

jmazanec15 · 2024-09-17T17:56:55Z

src/main/java/org/opensearch/knn/index/KNNSettings.java

@@ -66,6 +66,7 @@ public class KNNSettings {
     * Settings name
     */
    public static final String KNN_SPACE_TYPE = "index.knn.space_type";
+    public static final String INDEX_KNN_BUILD_VECTOR_DATA_STRUCTURE_THRESHOLD = "index.knn.build_vector_data_structure_threshold";


not a blocker, but I think we need to rename this setting. Need to brainstorm some ideas, but this is too verbose.

Some thoughts

index.knn.exact.threshold

index.ann.threshold

@VijayanB lets take this as AI to conclude on the naming.

… creation (opensearch-project#2007) * Fix native engine vector format test (opensearch-project#2103) Previosuly we were not creating hnsw file on segment flush for faiss engine. After successfully integrating hnsw file creation, we forgot to update unit test. Here, we will confirm that required files are being created based on field type. Signed-off-by: Vijayan Balasubramanian <[email protected]> * Introduce a setting to control whether to build graph Added new updatable index setting "build_vector_data_structure_threshold", which will be considered when to build braph or not for native engines. This is noop for lucene. This depends on use lucene format as prerequisite. We don't need to add flag since it is only enable if lucene format is already enabled. Signed-off-by: Vijayan Balasubramanian <[email protected]>

… creation (#2007) Added new updatable index setting "build_vector_data_structure_threshold", which will be considered when to build braph or not for native engines. This is noop for lucene. This depends on use lucene format as prerequisite. We don't need to add flag since it is only enable if lucene format is already enabled. Signed-off-by: Vijayan Balasubramanian <[email protected]>

… creation (opensearch-project#2007) Added new updatable index setting "build_vector_data_structure_threshold", which will be considered when to build braph or not for native engines. This is noop for lucene. This depends on use lucene format as prerequisite. We don't need to add flag since it is only enable if lucene format is already enabled. Signed-off-by: Vijayan Balasubramanian <[email protected]>

…t search when there are no engine files (#2188) * Introduce new setting to configure when to build graph during segment creation (#2007) Added new updatable index setting "build_vector_data_structure_threshold", which will be considered when to build braph or not for native engines. This is noop for lucene. This depends on use lucene format as prerequisite. We don't need to add flag since it is only enable if lucene format is already enabled. Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add integration test for binary vector values (#2142) Signed-off-by: Vijayan Balasubramanian <[email protected]> * Allow build graph greedily for quantization scenarios (#2175) Previosuly we only added support to build greedily for non quantization scenario. In this commit, we can remove that constraint, however, we cannot skip writing quanitization state since it is required irrespective of type of search is executed later. Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add exact search if no native engine files are available (#2136) * Add exact search if no engine files are in segments When graph is not available, plugin will return empty results. With this change, exact search will be performed when only no engine file is available in segment. We also don't need version check or feature flag because, option to not build vector data structure will only be available post 2.17. If an index is created using pre 2.17 version, segment will always have engine files and this feature will never be called during search. --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add support for radial search in exact search (#2174) * Add support for radial search in exact search When threshold value is set, knn plugin will not be creating graph. Hence, when search request is trigged during that time, exact search will return valid results. However, radial search was never included as part of exact search. This will break radial search when threshold is added and radial search is requested. In this commit, new method is introduced to accept min score and return documents that are greater than min score, similar to how radial search is performed by native engines. This search is independent of engine, but, radial search is supported only for FAISS engine out of all native engines. Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]>

…t search when there are no engine files (#2188) * Introduce new setting to configure when to build graph during segment creation (#2007) Added new updatable index setting "build_vector_data_structure_threshold", which will be considered when to build braph or not for native engines. This is noop for lucene. This depends on use lucene format as prerequisite. We don't need to add flag since it is only enable if lucene format is already enabled. Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add integration test for binary vector values (#2142) Signed-off-by: Vijayan Balasubramanian <[email protected]> * Allow build graph greedily for quantization scenarios (#2175) Previosuly we only added support to build greedily for non quantization scenario. In this commit, we can remove that constraint, however, we cannot skip writing quanitization state since it is required irrespective of type of search is executed later. Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add exact search if no native engine files are available (#2136) * Add exact search if no engine files are in segments When graph is not available, plugin will return empty results. With this change, exact search will be performed when only no engine file is available in segment. We also don't need version check or feature flag because, option to not build vector data structure will only be available post 2.17. If an index is created using pre 2.17 version, segment will always have engine files and this feature will never be called during search. --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add support for radial search in exact search (#2174) * Add support for radial search in exact search When threshold value is set, knn plugin will not be creating graph. Hence, when search request is trigged during that time, exact search will return valid results. However, radial search was never included as part of exact search. This will break radial search when threshold is added and radial search is requested. In this commit, new method is introduced to accept min score and return documents that are greater than min score, similar to how radial search is performed by native engines. This search is independent of engine, but, radial search is supported only for FAISS engine out of all native engines. Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 5a56829)

…nd perform exact search when there are no engine files (#2201) * Add support to build vector data structures greedily and perform exact search when there are no engine files (#2188) * Introduce new setting to configure when to build graph during segment creation (#2007) Added new updatable index setting "build_vector_data_structure_threshold", which will be considered when to build braph or not for native engines. This is noop for lucene. This depends on use lucene format as prerequisite. We don't need to add flag since it is only enable if lucene format is already enabled. Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add integration test for binary vector values (#2142) Signed-off-by: Vijayan Balasubramanian <[email protected]> * Allow build graph greedily for quantization scenarios (#2175) Previosuly we only added support to build greedily for non quantization scenario. In this commit, we can remove that constraint, however, we cannot skip writing quanitization state since it is required irrespective of type of search is executed later. Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add exact search if no native engine files are available (#2136) * Add exact search if no engine files are in segments When graph is not available, plugin will return empty results. With this change, exact search will be performed when only no engine file is available in segment. We also don't need version check or feature flag because, option to not build vector data structure will only be available post 2.17. If an index is created using pre 2.17 version, segment will always have engine files and this feature will never be called during search. --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> * Add support for radial search in exact search (#2174) * Add support for radial search in exact search When threshold value is set, knn plugin will not be creating graph. Hence, when search request is trigged during that time, exact search will return valid results. However, radial search was never included as part of exact search. This will break radial search when threshold is added and radial search is requested. In this commit, new method is introduced to accept min score and return documents that are greater than min score, similar to how radial search is performed by native engines. This search is independent of engine, but, radial search is supported only for FAISS engine out of all native engines. Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 5a56829) * Fix compilation issue due to package error Signed-off-by: Vijayan Balasubramanian <[email protected]> --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> Co-authored-by: Vijayan Balasubramanian <[email protected]>

VijayanB requested review from heemin32, navneet1v, vamshin, jmazanec15, naveentatikonda, junqiu-lei, martin-gaievski, ryanbogan and luyuncheng as code owners August 27, 2024 18:45

VijayanB changed the title ~~Introduce a setting to control whether to build graph or not~~ Introduce new setting to configure whether to build graph or not during segment creation Aug 27, 2024

navneet1v reviewed Aug 27, 2024

View reviewed changes

shatejas reviewed Aug 27, 2024

View reviewed changes

navneet1v reviewed Aug 28, 2024

View reviewed changes

navneet1v mentioned this pull request Aug 28, 2024

[FEATURE] Creating Vector data structures Greedily #1942

Closed

VijayanB force-pushed the build-graph-as-setting branch from d6a4a05 to 530dda8 Compare August 28, 2024 05:54

VijayanB changed the title ~~Introduce new setting to configure whether to build graph or not during segment creation~~ Introduce new setting to configure when to build graph during segment creation Sep 4, 2024

VijayanB changed the base branch from main to feature/build-vector-ds-greedily September 4, 2024 20:39

VijayanB force-pushed the build-graph-as-setting branch 2 times, most recently from db80d04 to d1cfae5 Compare September 4, 2024 20:46

VijayanB force-pushed the build-graph-as-setting branch from d1cfae5 to 5700d62 Compare September 4, 2024 21:48

VijayanB deleted the branch opensearch-project:feature/build-vector-ds-greedily September 4, 2024 21:50

VijayanB closed this Sep 4, 2024

VijayanB reopened this Sep 4, 2024

VijayanB closed this Sep 13, 2024

VijayanB reopened this Sep 13, 2024

VijayanB force-pushed the build-graph-as-setting branch from 39807c5 to f39ee0a Compare September 13, 2024 22:17

VijayanB force-pushed the build-graph-as-setting branch 5 times, most recently from b53e6d3 to 95b6b89 Compare September 16, 2024 19:24

shatejas reviewed Sep 16, 2024

View reviewed changes

VijayanB force-pushed the build-graph-as-setting branch from 95b6b89 to 32cff9b Compare September 16, 2024 20:48

Add unit test

4f5cc31

Signed-off-by: Vijayan Balasubramanian <[email protected]>

VijayanB force-pushed the build-graph-as-setting branch from 32cff9b to 4f5cc31 Compare September 17, 2024 00:35

shatejas approved these changes Sep 17, 2024

View reviewed changes

vibrantvarun reviewed Sep 17, 2024

View reviewed changes

jmazanec15 reviewed Sep 17, 2024

View reviewed changes

jmazanec15 approved these changes Sep 17, 2024

View reviewed changes

VijayanB merged commit c1c7040 into opensearch-project:feature/build-vector-ds-greedily Sep 17, 2024
27 of 28 checks passed

VijayanB mentioned this pull request Sep 28, 2024

Minor refactoring and refactored some unit test #2167

Merged

5 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Introduce new setting to configure when to build graph during segment creation #2007

Introduce new setting to configure when to build graph during segment creation #2007

VijayanB commented Aug 27, 2024 •

edited

Loading

navneet1v Aug 27, 2024

VijayanB Aug 27, 2024

navneet1v Aug 27, 2024

VijayanB Aug 27, 2024

navneet1v Aug 27, 2024

navneet1v Aug 27, 2024

shatejas Aug 27, 2024

navneet1v Aug 27, 2024

navneet1v Aug 27, 2024

navneet1v Aug 27, 2024

navneet1v Aug 27, 2024

jmazanec15 commented Aug 27, 2024

shatejas left a comment

shatejas Aug 27, 2024

heemin32 commented Aug 27, 2024 •

edited

Loading

navneet1v commented Aug 28, 2024 •

edited

Loading

navneet1v Aug 28, 2024

navneet1v Aug 28, 2024

navneet1v Aug 28, 2024

VijayanB Sep 4, 2024

VijayanB commented Aug 28, 2024

VijayanB commented Aug 28, 2024

VijayanB commented Sep 4, 2024

shatejas Sep 16, 2024

VijayanB Sep 16, 2024

shatejas Sep 16, 2024

VijayanB Sep 17, 2024

shatejas left a comment

vibrantvarun Sep 17, 2024

VijayanB Sep 17, 2024

jmazanec15 Sep 17, 2024

navneet1v Sep 17, 2024

VijayanB Sep 17, 2024

	public static final String KNN_BUILD_GRAPH = "index.knn.build_graph";
	public static final String KNN_BUILD_GRAPH = "index.knn.build_vector_datastructure";

	public NativeEngines990KnnVectorsFormat(final FlatVectorsFormat lucene99FlatVectorsFormat, boolean buildGraphSetting) {
	public NativeEngines990KnnVectorsFormat(final FlatVectorsFormat lucene99FlatVectorsFormat, boolean buildVectorDataStructure) {

Introduce new setting to configure when to build graph during segment creation #2007

Introduce new setting to configure when to build graph during segment creation #2007

Conversation

VijayanB commented Aug 27, 2024 • edited Loading

Description

Related Issues

Check List

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jmazanec15 commented Aug 27, 2024

shatejas left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

heemin32 commented Aug 27, 2024 • edited Loading

navneet1v commented Aug 28, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

VijayanB commented Aug 28, 2024

VijayanB commented Aug 28, 2024

VijayanB commented Sep 4, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

shatejas left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

VijayanB commented Aug 27, 2024 •

edited

Loading

heemin32 commented Aug 27, 2024 •

edited

Loading

navneet1v commented Aug 28, 2024 •

edited

Loading