[FLINK-36317][runtime] Populate ArchivedExecutionGraph with CheckpointStatsSnapshot in WaitingForResources state

[kaustubhbutte17]

What is the purpose of the change

When a Flink job fails and restarts, it transitions through Restarting → WaitingForResources. During this state, the previousExecutionGraph (which contains checkpoint statistics from the prior execution) is available, but the REST API/Web UI cannot access these stats because StateWithoutExecutionGraph.getJob() creates a sparse ArchivedExecutionGraph with empty checkpoint data.

This PR fixes that by:

1. Adding withCheckpointStatsSnapshot() to ArchivedExecutionGraph — creates a copy with different checkpoint stats
2. Overriding getJob() in WaitingForResources — attaches checkpoint stats from the previousExecutionGraph when available

Brief change log

• Added ArchivedExecutionGraph.withCheckpointStatsSnapshot() method
• Overrode WaitingForResources.getJob() to preserve checkpoint stats from the previous execution graph
• Added 2 unit tests in WaitingForResourcesTest

Verifying this change

This change is verified by new unit tests:

• testGetJobIncludesCheckpointStatsFromPreviousExecutionGraph — verifies checkpoint stats from a mock previous execution graph are preserved
• testGetJobWithoutPreviousExecutionGraphReturnsNullCheckpointStats — verifies the default behavior when no previous execution graph exists

Does this pull request potentially affect one of the following parts:

• Dependencies (does it add or upgrade a dependency): no
• The public API, i.e., is any changed class annotated with @Public(Evolving): no
• The serializers: no
• The runtime per-record code paths (performance sensitive): no
• Anything that affects deployment or recovery: no
• The (broadcasting) coordination functionality: no

Documentation

• Does this pull request introduce a new feature? no
• If yes, how is the feature documented? not applicable

[flinkbot]

CI report:

• 5692ff8 Azure: SUCCESS

Bot commands

The @flinkbot bot supports the following commands:

• @flinkbot run azure re-run the last Azure build

[spuru9]

Hi @kaustubhbutte17
Appears to be a linting issue mvn spotless:apply
Also can you add the component name in the PR title as specified in the guidelines/AGENTS.md.

[kaustubhbutte17]

Thanks @spuru9 for the review! Fixed both issues:

1. Ran mvn spotless:apply to fix formatting
2. Updated PR title to include [runtime] component

Pushed the fix in the latest commit.

[kaustubhbutte17]

@flinkbot run azure

[kaustubhbutte17]

Gentle ping @spuru9 , or let me know if there is some other way to contact someone for diff reviews, this is my first time contributing

[hiSandog]

The approach looks sound — overriding getJob() in WaitingForResources to carry checkpoint stats from previousExecutionGraph into the archived graph is a clean, localized fix. The withCheckpointStatsSnapshot copy method on ArchivedExecutionGraph is straightforward and avoids mutating shared state.

One minor observation: the second test (testGetJobWithoutPreviousExecutionGraphReturnsNullCheckpointStats) only asserts that archivedGraph is not null, but doesn't explicitly verify that getCheckpointStatsSnapshot() is null. Adding an explicit assertThat(archivedGraph.getCheckpointStatsSnapshot()).isNull() would make the negative case more robust against future refactors.

CI is green and the change is well-scoped — nice first contribution!