Implement list and delete snapshot functionality in Python SDK by shrutiyam-glitch · Pull Request #448 · kubernetes-sigs/agent-sandbox

shrutiyam-glitch · 2026-03-20T15:54:34Z

This PR implements the ability to list and delete Pod Snapshots within the PodSnapshotSandboxClient.

Core Logic Implementation:
* list_snapshots(grouping_labels, ready_only):
-- Fetches PodSnapshots using Kubernetes label selectors.
-- Returns a list of snapshots sorted by creation timestamp (newest first).
-- Supports filtering for snapshots in the Ready state.
*delete_snapshots(grouping_labels, snapshot_uid):
-- Allows deleting a specific snapshot by UID.
-- Supports bulk deletion based on grouping labels or deleting all snapshots associated with the pod.

Testing Done:

Integration Test: Added test_podsnapshot_extension.py which verifies the full E2E flow:
-- creating multiple snapshots, listing them, and performing a cleanup deletion
Unit tests are added

Output:

Integration Test (clients/python/agentic-sandbox-client/test_podsnapshot_extension.py):

$ python3 test_podsnapshot_extension.py --template-name python-counter-template --namespace sandbox-test
--- Starting Sandbox Client Test (Namespace: sandbox-test, Port: 8888) ---

***** Phase 1: Starting Counter *****

======= Testing Pod Snapshot Extension =======
Creating first pod snapshot 'test-snapshot-10' after 10 seconds...
Trigger Name: test-snapshot-10-20260320-075957-3f239b0f
Snapshot UID: fdb1a93a-37ca-4ac2-b339-3cf2025b2eb0
Success: True
Error Code: 0
Error Reason: 

Creating second pod snapshot 'test-snapshot-20' after 10 seconds...
Trigger Name: test-snapshot-20-20260320-080011-b3a7cf3d
Snapshot UID: ea7b86eb-2813-430e-bd2e-95fac639a720
Success: True
Error Code: 0
Error Reason: 
Recent snapshot UID: ea7b86eb-2813-430e-bd2e-95fac639a720

***** List all existing ready snapshots associated with the sandbox. *****
Snapshot ID: ea7b86eb-2813-430e-bd2e-95fac639a720, Source Pod: sandbox-claim-79aee963, Creation Time: 2026-03-20T08:00:11Z
Snapshot ID: fdb1a93a-37ca-4ac2-b339-3cf2025b2eb0, Source Pod: sandbox-claim-79aee963, Creation Time: 2026-03-20T07:59:58Z

***** Phase 2: Restoring from most recent snapshot & Verifying *****
Pod was restored from the most recent snapshot.

**** Deleting snapshots *****
Deleted Snapshots: ['ea7b86eb-2813-430e-bd2e-95fac639a720', 'fdb1a93a-37ca-4ac2-b339-3cf2025b2eb0']
--- Pod Snapshot Test Passed! ---

--- Sandbox Client Test Finished ---

netlify · 2026-03-20T15:54:41Z

✅ Deploy Preview for agent-sandbox canceled.

Name	Link
🔨 Latest commit	`ed5d1de`
🔍 Latest deploy log	https://app.netlify.com/projects/agent-sandbox/deploys/69bd6dbe489b420008fc8465

k8s-ci-robot · 2026-03-20T15:54:41Z

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: shrutiyam-glitch
Once this PR has been reviewed and has the lgtm label, please assign justinsb for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

clients/python/agentic-sandbox-client/OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

vicentefb · 2026-03-20T17:41:13Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+    """Result of a list snapshots operation."""
+
+    success: bool
+    snapshots: list[dict[str, str]]


Instead of returning a list[dict[str, str]], wdyt about creating a dedicated dataclass (e.g., SnapshotInfo or SnapshotDetail) for the snapshot representation ?

vicentefb · 2026-03-20T17:42:40Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+
+            valid_snapshots.append(
+                {
+                    "snapshot_id": metadata.get("name"),


We are mapping metadata.get("name") to snapshot_id here. However, down below in delete_snapshots, the parameter is named snapshot_uid. Please add a small docstring note or standardizing the terminology across the client to snapshot_name to prevent confusion.

vicentefb · 2026-03-20T17:43:19Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+            )
+
+        # Sort snapshots by creation timestamp descending
+        valid_snapshots.sort(key=lambda x: x["creationTimestamp"], reverse=True)


If a snapshot object is missing creationTimestamp or it explicitly evaluates to None for any unforeseen reason, sorting might throw a TypeError. To be safe, you can use key=lambda x: x["creationTimestamp"] or "".

vicentefb · 2026-03-20T17:44:23Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+        - If snapshot_uid is provided, deletes that specific snapshot.
+        - If grouping_labels is provided, deletes all snapshots matching the grouping labels.
+        - If not provided, deletes ALL snapshots for this pod.
+        Returns a DeleteSnapshotResult containing the list of successfully deleted snapshots.


Just to confirm my understanding here. If a user passes both snapshot_uid and grouping_labels to this method, grouping_labels will be silently ignored, right ? It would be safer to either explicitly enforce mutual exclusivity by raising a ValueError, or at least log a warning that the labels are being ignored when a UID is specified or document this case...

vicentefb · 2026-03-20T17:45:02Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+            snapshots_to_delete.append(snapshot_uid)
+        else:
+            logger.info(
+                "No snapshot_uid provided. Deleting ALL snapshots for this pod."


This log message says "Deleting ALL snapshots for this pod." But, if grouping_labels is provided, it's actually filtering and only deleting a matching subset. Should we reword this to something like: "No snapshot_uid provided. Fetching snapshots to delete based on pod name and labels."

vicentefb · 2026-03-20T17:45:55Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+                    error_code=SNAPSHOT_ERROR_CODE,
+                )
+            if (
+                snapshots_result


snapshots_result is a ListSnapshotResult dataclass instance, so shouldn't this always be "true" ? . Maybe you can simplify this condition to:
if snapshots_result.success and snapshots_result.snapshots:

vicentefb · 2026-03-20T17:46:40Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+                    name=uid,
+                )
+                logger.info(f"PodSnapshot '{uid}' deleted.")
+                deleted_snapshots.append(uid)


When deleting a list of snapshots (batch deletion), if one deletion throws an ApiException (other than 404), the function returns immediately. This leaves the remaining targeted snapshots undeleted. Let's accumulate the errors in a list and continuing the loop, then returning an aggregated failure response at the end so it attempts to delete as many as they were intended.

If one deletion fails in the middle, there is another case where it returns success=False, but partial delete happened

tomergee · 2026-03-21T06:11:03Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+                and snapshots_result.snapshots
+            ):
+                snapshots_to_delete = [
+                    s["snapshot_id"] for s in snapshots_result.snapshots


is it the snapshot name or uid here? not clear, if the name is needed Rename the parameter to snapshot_name, could be confusing

tomergee · 2026-03-21T06:15:30Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+                )
+            if (
+                snapshots_result
+                and snapshots_result.success


You already checked not snapshots_result.success above

tomergee · 2026-03-21T06:22:29Z

...ts/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/test_podsnapshot_client.py

@@ -19,11 +19,13 @@
    PodSnapshotSandboxClient,


Missing Tests
test for list_snapshots with ready_only=False

There is no unit test verifying that passing ready_only=False includes non-ready snapshots. This is a key feature path.

test for delete_snapshots with snapshot_uid and grouping_labels both provided

The docstring implies only snapshot_uid is used when both are provided — but there's no test asserting that grouping_labels is ignored in that case.

test for delete_snapshots with a generic Exception

delete_snapshots loop catches Exception in addition to ApiException, but there's no unit test for this code path.

tomergee · 2026-03-21T06:24:54Z

clients/python/agentic-sandbox-client/k8s_agent_sandbox/gke_extensions/podsnapshot_client.py

+
+        # Sort snapshots by creation timestamp descending
+        valid_snapshots.sort(key=lambda x: x["creationTimestamp"], reverse=True)
+        logger.info(f"Found {len(valid_snapshots)} ready snapshots.")


This will print also when ready_only=False, just change the wording

tomergee · 2026-03-21T06:26:23Z

clients/python/agentic-sandbox-client/test_podsnapshot_extension.py

+                    )
+            else:
+                print(
+                    f"No ready snapshots found or failed: {list_result.error_reason if list_result else ''}"


Is this check complete?, it just check the object type, will not reach else

shrutiyam-glitch added 3 commits March 20, 2026 08:27

PSS list delete

24edd83

Merge branch 'main' into pss-list-delete

a33bc51

Update unit test

ed5d1de

k8s-ci-robot requested review from igooch and justinsb March 20, 2026 15:54

k8s-ci-robot added cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. size/L Denotes a PR that changes 100-499 lines, ignoring generated files. labels Mar 20, 2026

vicentefb reviewed Mar 20, 2026

View reviewed changes

tomergee reviewed Mar 21, 2026

View reviewed changes

Conversation

shrutiyam-glitch commented Mar 20, 2026

Uh oh!

netlify bot commented Mar 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

✅ Deploy Preview for agent-sandbox canceled.

Uh oh!

k8s-ci-robot commented Mar 20, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tomergee Mar 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

netlify bot commented Mar 20, 2026 •

edited

Loading

tomergee Mar 21, 2026 •

edited

Loading