Add more initializer-related info to /insights API (#20572)

* [ws-manager, ws-daemon] Store initializer metrics in workspace.Status.InitializerMetrics

Tool: gitpod/catfood.gitpod.cloud

* [ws-mananger-api, -mk2] Emit new field .Status.InitializerMetrics

Tool: gitpod/catfood.gitpod.cloud

* [db] Introduce DBWorkspaceInstanceMetrics and persist all metrics from ws-manager-api into it

Tool: gitpod/catfood.gitpod.cloud

* [api] Expose session.Metrics.InitializerMetrics

Tool: gitpod/catfood.gitpod.cloud

* [dashboard] Export metrics into CSV

Tool: gitpod/catfood.gitpod.cloud

* [content-service] Fix: emit fromBackup stats

Tool: gitpod/catfood.gitpod.cloud

* Update components/ws-manager-api/core.proto

Co-authored-by: Filip Troníček <[email protected]>

---------

Co-authored-by: Filip Troníček <[email protected]>

Author: geropl

components/content-service/pkg/initializer/initializer.go

@@ -468,20 +468,38 @@ func InitializeWorkspace(ctx context.Context, location string, remoteStorage sto
 		}
 	}
 
-	// Run the initializer
+	// Try to download a backup first
+	initialSize, fsErr := getFsUsage()
+	if fsErr != nil {
+		log.WithError(fsErr).Error("could not get disk usage")
+	}
+	downloadStart := time.Now()
 	hasBackup, err := remoteStorage.Download(ctx, location, storage.DefaultBackup, cfg.mappings)
 	if err != nil {
 		return src, nil, xerrors.Errorf("cannot restore backup: %w", err)
 	}
+	downloadDuration := time.Since(downloadStart)
 
 	span.SetTag("hasBackup", hasBackup)
 	if hasBackup {
 		src = csapi.WorkspaceInitFromBackup
-	} else {
-		src, stats, err = cfg.Initializer.Run(ctx, cfg.mappings)
-		if err != nil {
-			return src, nil, xerrors.Errorf("cannot initialize workspace: %w", err)
+
+		currentSize, fsErr := getFsUsage()
+		if fsErr != nil {
+			log.WithError(fsErr).Error("could not get disk usage")
 		}
+		stats = []csapi.InitializerMetric{{
+			Type:     "fromBackup",
+			Duration: downloadDuration,
+			Size:     currentSize - initialSize,
+		}}
+		return
+	}
+
+	// If there is not backup, run the initializer
+	src, stats, err = cfg.Initializer.Run(ctx, cfg.mappings)
+	if err != nil {
+		return src, nil, xerrors.Errorf("cannot initialize workspace: %w", err)
 	}
 
 	return

components/dashboard/src/Insights.tsx

@@ -63,7 +63,7 @@ export const Insights = () => {
                         "md:flex-row md:items-center md:space-x-4 md:space-y-0",
                     )}
                 >
-                    <DownloadUsage to={toDate} disabled={isLackingPermissions} />
+                    <DownloadInsights to={toDate} disabled={isLackingPermissions} />
                 </div>
 
                 <div
@@ -166,7 +166,7 @@ type DownloadUsageProps = {
     to: Timestamp;
     disabled?: boolean;
 };
-export const DownloadUsage = ({ to, disabled }: DownloadUsageProps) => {
+export const DownloadInsights = ({ to, disabled }: DownloadUsageProps) => {
     const { data: org } = useCurrentOrg();
     const { toast } = useToast();
     // When we start the download, we disable the button for a short time

components/dashboard/src/insights/download/download-sessions.ts

@@ -15,7 +15,7 @@ import dayjs from "dayjs";
 import { useQuery, useQueryClient } from "@tanstack/react-query";
 import { useCallback } from "react";
 import { noPersistence } from "../../data/setup";
-import { Timestamp } from "@bufbuild/protobuf";
+import { Duration, Timestamp } from "@bufbuild/protobuf";
 
 const pageSize = 100;
 const maxPages = 100; // safety limit if something goes wrong with pagination
@@ -154,6 +154,16 @@ const displayTime = (timestamp?: Timestamp) => {
     return timestamp.toDate().toISOString();
 };
 
+const renderDuration = (duration?: Duration): string => {
+    if (!duration) {
+        return "";
+    }
+
+    let seconds = Number(duration.seconds);
+    seconds += duration.nanos / 1_000_000_000;
+    return seconds.toString(10);
+};
+
 export const transformSessionRecord = (session: WorkspaceSession) => {
     const initializerType = session.workspace?.spec?.initializer?.specs;
     const prebuildInitializer = initializerType?.find((i) => i.spec.case === "prebuild")?.spec.value as
@@ -190,6 +200,20 @@ export const transformSessionRecord = (session: WorkspaceSession) => {
         timeout: session.workspace?.spec?.timeout?.inactivity?.seconds,
         editor: session.workspace?.spec?.editor?.name,
         editorVersion: session.workspace?.spec?.editor?.version, // indicates whether user selected the stable or latest editor release channel
+
+        // initializer metrics
+        contentInitGitDuration: renderDuration(session.metrics?.initializerMetrics?.git?.duration),
+        contentInitGitSize: session.metrics?.initializerMetrics?.git?.size,
+        contentInitFileDownloadDuration: renderDuration(session.metrics?.initializerMetrics?.fileDownload?.duration),
+        contentInitFileDownloadSize: session.metrics?.initializerMetrics?.fileDownload?.size,
+        contentInitSnapshotDuration: renderDuration(session.metrics?.initializerMetrics?.snapshot?.duration),
+        contentInitSnapshotSize: session.metrics?.initializerMetrics?.snapshot?.size,
+        contentInitBackupDuration: renderDuration(session.metrics?.initializerMetrics?.backup?.duration),
+        contentInitBackupSize: session.metrics?.initializerMetrics?.backup?.size,
+        contentInitPrebuildDuration: renderDuration(session.metrics?.initializerMetrics?.prebuild?.duration),
+        contentInitPrebuildSize: session.metrics?.initializerMetrics?.prebuild?.size,
+        contentInitCompositeDuration: renderDuration(session.metrics?.initializerMetrics?.composite?.duration),
+        contentInitCompositeSize: session.metrics?.initializerMetrics?.composite?.size,
     };
 
     return row;

components/gitpod-db/src/typeorm/entity/db-workspace-instance-metrics.ts

@@ -0,0 +1,20 @@
+/**
+ * Copyright (c) 2025 Gitpod GmbH. All rights reserved.
+ * Licensed under the GNU Affero General Public License (AGPL).
+ * See License.AGPL.txt in the project root for license information.
+ */
+
+import { Entity, Column, PrimaryColumn } from "typeorm";
+import { WorkspaceInstanceMetrics } from "@gitpod/gitpod-protocol";
+
+@Entity()
+export class DBWorkspaceInstanceMetrics {
+    @PrimaryColumn()
+    instanceId: string;
+
+    @Column("json", { nullable: true })
+    metrics?: WorkspaceInstanceMetrics;
+
+    @Column()
+    _lastModified: Date;
+}

components/gitpod-db/src/typeorm/migration/1739892121734-AddWorkspaceInstanceMetricsTable.ts

@@ -0,0 +1,23 @@
+/**
+ * Copyright (c) 2025 Gitpod GmbH. All rights reserved.
+ * Licensed under the GNU Affero General Public License (AGPL).
+ * See License.AGPL.txt in the project root for license information.
+ */
+
+import { MigrationInterface, QueryRunner } from "typeorm";
+
+export class AddWorkspaceInstanceMetricsTable1739892121734 implements MigrationInterface {
+    public async up(queryRunner: QueryRunner): Promise<any> {
+        await queryRunner.query(`CREATE TABLE IF NOT EXISTS d_b_workspace_instance_metrics (
+            instanceId char(36) NOT NULL,
+            metrics JSON,
+            _lastModified timestamp(6) NOT NULL DEFAULT CURRENT_TIMESTAMP(6) ON UPDATE CURRENT_TIMESTAMP(6),
+            PRIMARY KEY (instanceId),
+            KEY ind_dbsync (_lastModified)
+        ) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;`);
+    }
+
+    public async down(queryRunner: QueryRunner): Promise<any> {
+        await queryRunner.query(`DROP TABLE IF EXISTS d_b_workspace_instance_metrics`);
+    }
+}

components/gitpod-db/src/typeorm/workspace-db-impl.ts

@@ -19,6 +19,7 @@ import {
     WorkspaceAndInstance,
     WorkspaceInfo,
     WorkspaceInstance,
+    WorkspaceInstanceMetrics,
     WorkspaceInstanceUser,
     WorkspaceSession,
     WorkspaceType,
@@ -62,6 +63,7 @@ import { TypeORM } from "./typeorm";
 import { ApplicationError, ErrorCodes } from "@gitpod/gitpod-protocol/lib/messaging/error";
 import { DBProject } from "./entity/db-project";
 import { PrebuiltWorkspaceWithWorkspace } from "@gitpod/gitpod-protocol/src/protocol";
+import { DBWorkspaceInstanceMetrics } from "./entity/db-workspace-instance-metrics";
 
 type RawTo<T> = (instance: WorkspaceInstance, ws: Workspace) => T;
 interface OrderBy {
@@ -109,6 +111,10 @@ export class TypeORMWorkspaceDBImpl extends TransactionalDBImpl<WorkspaceDB> imp
         );
     }
 
+    private async getWorkspaceInstanceMetricsRepo(): Promise<Repository<DBWorkspaceInstanceMetrics>> {
+        return (await this.getEntityManager()).getRepository<DBWorkspaceInstanceMetrics>(DBWorkspaceInstanceMetrics);
+    }
+
     public async connect(maxTries: number = 3, timeout: number = 2000): Promise<void> {
         let tries = 1;
         while (tries <= maxTries) {
@@ -459,27 +465,39 @@ export class TypeORMWorkspaceDBImpl extends TransactionalDBImpl<WorkspaceDB> imp
         offset: number,
     ): Promise<WorkspaceSession[]> {
         const workspaceInstanceRepo = await this.getWorkspaceInstanceRepo();
+
         // The query basically selects all workspace instances for the given owner, whose startDate is within the period, and which are either:
         //  - not stopped yet, or
         //  - is stopped or stopping.
-        const sessions = await workspaceInstanceRepo
+        type JoinResult = DBWorkspaceInstance & {
+            metrics: DBWorkspaceInstanceMetrics | undefined;
+            workspace: DBWorkspace;
+        };
+        const sessions = (await workspaceInstanceRepo
             .createQueryBuilder("wsi")
             .leftJoinAndMapOne("wsi.workspace", DBWorkspace, "ws", "ws.id = wsi.workspaceId")
+            .leftJoinAndMapOne("wsi.metrics", DBWorkspaceInstanceMetrics, "wsim", "wsim.instanceId = wsi.id")
             .where("ws.organizationId = :organizationId", { organizationId })
             .andWhere("wsi.creationTime >= :periodStart", { periodStart: periodStart.toISOString() })
             .andWhere("wsi.creationTime <= :periodEnd", { periodEnd: periodEnd.toISOString() })
             .orderBy("wsi.creationTime", "DESC")
             .skip(offset)
             .take(limit)
-            .getMany();
+            .getMany()) as JoinResult[];
 
-        const resultSessions: { instance: WorkspaceInstance; workspace: Workspace }[] = [];
+        const resultSessions: {
+            instance: WorkspaceInstance;
+            workspace: Workspace;
+            metrics?: WorkspaceInstanceMetrics;
+        }[] = [];
         for (const session of sessions) {
             resultSessions.push({
-                workspace: (session as any).workspace,
+                workspace: session.workspace,
                 instance: session,
+                metrics: session.metrics?.metrics,
             });
             delete (session as any).workspace;
+            delete (session as any).metrics;
         }
         return resultSessions;
     }
@@ -1143,6 +1161,36 @@ export class TypeORMWorkspaceDBImpl extends TransactionalDBImpl<WorkspaceDB> imp
         const res = await query.getMany();
         return res.map((r) => r.info);
     }
+
+    async storeMetrics(instanceId: string, metrics: WorkspaceInstanceMetrics): Promise<WorkspaceInstanceMetrics> {
+        const repo = await this.getWorkspaceInstanceMetricsRepo();
+        const result = await repo.save({
+            instanceId,
+            metrics,
+        });
+        return result.metrics;
+    }
+
+    async getMetrics(instanceId: string): Promise<WorkspaceInstanceMetrics | undefined> {
+        const repo = await this.getWorkspaceInstanceMetricsRepo();
+        const dbMetrics = await repo.findOne({ where: { instanceId } });
+        return dbMetrics?.metrics;
+    }
+
+    async updateMetrics(
+        instanceId: string,
+        update: WorkspaceInstanceMetrics,
+        merge: (current: WorkspaceInstanceMetrics, update: WorkspaceInstanceMetrics) => WorkspaceInstanceMetrics,
+    ): Promise<WorkspaceInstanceMetrics> {
+        return await this.transaction(async (db) => {
+            const current = await db.getMetrics(instanceId);
+            if (!current) {
+                return await db.storeMetrics(instanceId, update);
+            }
+            const merged = merge(current, update);
+            return await db.storeMetrics(instanceId, merged);
+        });
+    }
 }
 
 type InstanceJoinResult = DBWorkspace & { instance: WorkspaceInstance };

components/gitpod-db/src/workspace-db.ts

@@ -23,6 +23,7 @@ import {
     WorkspaceSession,
     PrebuiltWorkspaceWithWorkspace,
     PrebuildWithStatus,
+    WorkspaceInstanceMetrics,
 } from "@gitpod/gitpod-protocol";
 
 export type MaybeWorkspace = Workspace | undefined;
@@ -196,4 +197,12 @@ export interface WorkspaceDB {
 
     storePrebuildInfo(prebuildInfo: PrebuildInfo): Promise<void>;
     findPrebuildInfos(prebuildIds: string[]): Promise<PrebuildInfo[]>;
+
+    storeMetrics(instanceId: string, metrics: WorkspaceInstanceMetrics): Promise<WorkspaceInstanceMetrics>;
+    getMetrics(instanceId: string): Promise<WorkspaceInstanceMetrics | undefined>;
+    updateMetrics(
+        instanceId: string,
+        update: WorkspaceInstanceMetrics,
+        merge: (current: WorkspaceInstanceMetrics, update: WorkspaceInstanceMetrics) => WorkspaceInstanceMetrics,
+    ): Promise<WorkspaceInstanceMetrics>;
 }

components/gitpod-protocol/src/protocol.ts

@@ -4,7 +4,7 @@
  * See License.AGPL.txt in the project root for license information.
  */
 
-import { WorkspaceInstance, PortVisibility, PortProtocol } from "./workspace-instance";
+import { WorkspaceInstance, PortVisibility, PortProtocol, WorkspaceInstanceMetrics } from "./workspace-instance";
 import { RoleOrPermission } from "./permission";
 import { Project } from "./teams-projects-protocol";
 import { createHash } from "crypto";
@@ -1390,6 +1390,7 @@ export namespace WorkspaceInstancePortsChangedEvent {
 export interface WorkspaceSession {
     workspace: Workspace;
     instance: WorkspaceInstance;
+    metrics?: WorkspaceInstanceMetrics;
 }
 export interface WorkspaceInfo {
     workspace: Workspace;

components/gitpod-protocol/src/workspace-instance.ts

@@ -332,14 +332,43 @@ export interface ImageBuildLogInfo {
  * Holds metrics about the workspace instance
  */
 export interface WorkspaceInstanceMetrics {
-    image?: Partial<{
-        /**
-         * the total size of the image in bytes (includes Gitpod-specific layers like IDE)
-         */
-        totalSize: number;
-        /**
-         * the size of the workspace image in bytes
-         */
-        workspaceImageSize: number;
-    }>;
+    image?: ImageMetrics;
+
+    /**
+     * Metrics about the workspace initializer
+     */
+    initializerMetrics?: InitializerMetrics;
+}
+
+export interface ImageMetrics {
+    /**
+     * the total size of the image in bytes (includes Gitpod-specific layers like IDE)
+     */
+    totalSize?: number;
+
+    /**
+     * the size of the workspace image in bytes
+     */
+    workspaceImageSize?: number;
+}
+
+export interface InitializerMetrics {
+    git?: InitializerMetric;
+    fileDownload?: InitializerMetric;
+    snapshot?: InitializerMetric;
+    backup?: InitializerMetric;
+    prebuild?: InitializerMetric;
+    composite?: InitializerMetric;
+}
+
+export interface InitializerMetric {
+    /**
+     * Duration in milliseconds
+     */
+    duration: number;
+
+    /**
+     * Size in bytes
+     */
+    size: number;
 }