fix(backend): Continue stats accounting on aborted or broken executions (#9921)

This is a follow-up to
https://github.com/Significant-Gravitas/AutoGPT/pull/9903

The continued graph execution restarted all the execution stats from
zero, making the execution stats misleading.

### Changes 🏗️

Continue the execution stats when continuing the graph execution.

### Checklist 📋

#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
  <!-- Put your test plan here: -->
- [x] Existing tests, manual graph run with the graph execution aborted
midway.

autogpt_platform/backend/backend/data/execution.py

@@ -30,7 +30,7 @@ from prisma.types import (
     AgentNodeExecutionUpdateInput,
     AgentNodeExecutionWhereInput,
 )
-from pydantic import BaseModel
+from pydantic import BaseModel, ConfigDict
 from pydantic.fields import Field
 
 from backend.server.v2.store.exceptions import DatabaseError
@@ -69,10 +69,55 @@ class GraphExecutionMeta(BaseDbModel):
     ended_at: datetime
 
     class Stats(BaseModel):
-        cost: int = Field(..., description="Execution cost (cents)")
-        duration: float = Field(..., description="Seconds from start to end of run")
-        node_exec_time: float = Field(..., description="Seconds of total node runtime")
-        node_exec_count: int = Field(..., description="Number of node executions")
+        model_config = ConfigDict(
+            extra="allow",
+            arbitrary_types_allowed=True,
+        )
+
+        cost: int = Field(
+            default=0,
+            description="Execution cost (cents)",
+        )
+        duration: float = Field(
+            default=0,
+            description="Seconds from start to end of run",
+        )
+        duration_cpu_only: float = Field(
+            default=0,
+            description="CPU sec of duration",
+        )
+        node_exec_time: float = Field(
+            default=0,
+            description="Seconds of total node runtime",
+        )
+        node_exec_time_cpu_only: float = Field(
+            default=0,
+            description="CPU sec of node_exec_time",
+        )
+        node_exec_count: int = Field(
+            default=0,
+            description="Number of node executions",
+        )
+        node_error_count: int = Field(
+            default=0,
+            description="Number of node errors",
+        )
+        error: str | None = Field(
+            default=None,
+            description="Error message if any",
+        )
+
+        def to_db(self) -> GraphExecutionStats:
+            return GraphExecutionStats(
+                cost=self.cost,
+                walltime=self.duration,
+                cputime=self.duration_cpu_only,
+                nodes_walltime=self.node_exec_time,
+                nodes_cputime=self.node_exec_time_cpu_only,
+                node_count=self.node_exec_count,
+                node_error_count=self.node_error_count,
+                error=self.error,
+            )
 
     stats: Stats | None
 
@@ -106,8 +151,16 @@ class GraphExecutionMeta(BaseDbModel):
                 GraphExecutionMeta.Stats(
                     cost=stats.cost,
                     duration=stats.walltime,
+                    duration_cpu_only=stats.cputime,
                     node_exec_time=stats.nodes_walltime,
+                    node_exec_time_cpu_only=stats.nodes_cputime,
                     node_exec_count=stats.node_count,
+                    node_error_count=stats.node_error_count,
+                    error=(
+                        str(stats.error)
+                        if isinstance(stats.error, Exception)
+                        else stats.error
+                    ),
                 )
                 if stats
                 else None

autogpt_platform/backend/backend/executor/manager.py

@@ -633,7 +633,12 @@ class Executor:
             return
 
         timing_info, (exec_stats, status, error) = cls._on_graph_execution(
-            graph_exec, cancel, log_metadata
+            graph_exec=graph_exec,
+            cancel=cancel,
+            log_metadata=log_metadata,
+            execution_stats=(
+                exec_meta.stats.to_db() if exec_meta.stats else GraphExecutionStats()
+            ),
         )
         exec_stats.walltime = timing_info.wall_time
         exec_stats.cputime = timing_info.cpu_time
@@ -704,6 +709,7 @@ class Executor:
         graph_exec: GraphExecutionEntry,
         cancel: threading.Event,
         log_metadata: LogMetadata,
+        execution_stats: GraphExecutionStats,
     ) -> tuple[GraphExecutionStats, ExecutionStatus, Exception | None]:
         """
         Returns:
@@ -711,7 +717,6 @@ class Executor:
             ExecutionStatus: The final status of the graph execution.
             Exception | None: The error that occurred during the execution, if any.
         """
-        execution_stats = GraphExecutionStats()
         execution_status = ExecutionStatus.RUNNING
         error = None
         finished = False

autogpt_platform/frontend/src/lib/autogpt-server-api/types.ts

@@ -282,7 +282,9 @@ export type GraphExecutionMeta = {
   stats?: {
     cost: number;
     duration: number;
+    duration_cpu_only: number;
     node_exec_time: number;
+    node_exec_time_cpu_only: number;
     node_exec_count: number;
   };
 };