diff --git a/spark-plugin/plugin/src/main/scala/org/apache/spark/dataflint/listener/DataflintStore.scala b/spark-plugin/plugin/src/main/scala/org/apache/spark/dataflint/listener/DataflintStore.scala
index 0aca308..121be8a 100644
--- a/spark-plugin/plugin/src/main/scala/org/apache/spark/dataflint/listener/DataflintStore.scala
+++ b/spark-plugin/plugin/src/main/scala/org/apache/spark/dataflint/listener/DataflintStore.scala
@@ -5,6 +5,6 @@ import org.apache.spark.util.kvstore.KVStore
 
 class DataflintStore(val store: KVStore) {
   def icebergCommits(offset: Int, length: Int): Seq[IcebergCommitInfo] = {
-    KVUtils.mapToSeq(store.view(classOf[IcebergCommitWrapper]).skip(offset).max(length))(_.info).sortBy(_.executionId)
+    KVUtils.mapToSeq(store.view(classOf[IcebergCommitWrapper]))(_.info).filter(_.executionId >= offset).take(length).sortBy(_.executionId)
   }
 }
diff --git a/spark-ui/src/reducers/SqlReducer.ts b/spark-ui/src/reducers/SqlReducer.ts
index dc7304f..c02b1a3 100644
--- a/spark-ui/src/reducers/SqlReducer.ts
+++ b/spark-ui/src/reducers/SqlReducer.ts
@@ -340,7 +340,7 @@ export function calculateSqlStore(
     // case 1: SQL does not exist, we add it
     if (currentSql === undefined) {
       updatedSqls.push(calculateSql(newSql, plan, icebergCommit));
-      // From here currentSql must not be null, and currentSql can't be COMPLETED as it would not be requested by API
+      // From here currentSql must not be null
       // case 2: plan status changed from running to completed, so we need to update the SQL
     } else if (
       newSql.status === SqlStatus.Completed.valueOf() ||
diff --git a/spark-ui/src/services/SparkApi.tsx b/spark-ui/src/services/SparkApi.tsx
index 66cde10..069bc9d 100644
--- a/spark-ui/src/services/SparkApi.tsx
+++ b/spark-ui/src/services/SparkApi.tsx
@@ -21,12 +21,19 @@ import {
   updateDuration,
 } from "../reducers/SparkSlice";
 import { AppDispatch } from "../Store";
+import { timeStrToEpocTime } from "../utils/FormatUtils";
 import { IS_HISTORY_SERVER_MODE } from "../utils/UrlConsts";
 import { isDataFlintSaaSUI } from "../utils/UrlUtils";
 import { MixpanelService } from "./MixpanelService";
 
 const POLL_TIME = 1000;
 const SQL_QUERY_LENGTH = 1000;
+const SQL_LOOKBACK_QUERY_TIME = 5000;
+
+type SQLFinishTime = {
+  id: number;
+  finishTime: number;
+};
 
 class SparkAPI {
   basePath: string;
@@ -39,10 +46,23 @@ class SparkAPI {
   applicationsPath: string;
   dispatch: AppDispatch;
   lastCompletedSqlId: number = -1;
+  sqlIdToFinishTime: Record<number, SQLFinishTime> = {};
   pollingStopped: boolean = false;
   historyServerMode: boolean = false;
   icebergEnabled: boolean = false;
 
+  private findSqlIdToQueryFrom(): number {
+    const currentTime = Date.now();
+    const ids = Object.keys(this.sqlIdToFinishTime).map(id => parseInt(id));
+    const lookbackSqls = ids.filter(id => this.sqlIdToFinishTime[id] !== null && this.sqlIdToFinishTime[id].finishTime + SQL_LOOKBACK_QUERY_TIME > currentTime);
+
+    if (lookbackSqls.length > 0) {
+      return Math.min(...lookbackSqls);
+    }
+
+    return this.lastCompletedSqlId + 1;
+  }
+
   private get applicationPath(): string {
     return (
       `${this.apiPath}/applications/${this.appId}` +
@@ -241,15 +261,16 @@ class SparkAPI {
       const sparkJobs: SparkJobs = await this.queryData(this.jobsPath);
       this.dispatch(setSparkJobs({ value: sparkJobs }));
 
+      const sqlIdToQueryFrom = this.findSqlIdToQueryFrom();
       const sparkSQLs: SparkSQLs = await this.queryData(
-        this.buildSqlPath(this.lastCompletedSqlId + 1),
+        this.buildSqlPath(sqlIdToQueryFrom),
       );
       const sparkPlans: SQLPlans = await this.queryData(
-        this.buildSqlPlanPath(this.lastCompletedSqlId + 1),
+        this.buildSqlPlanPath(sqlIdToQueryFrom),
       );
       let icebergInfo: IcebergInfo = { commitsInfo: [] };
       if (this.icebergEnabled) {
-        icebergInfo = await this.queryData(this.buildIcebergPath(this.lastCompletedSqlId + 1));
+        icebergInfo = await this.queryData(this.buildIcebergPath(sqlIdToQueryFrom));
       }
 
       if (sparkSQLs.length !== 0) {
@@ -264,9 +285,13 @@ class SparkAPI {
         if (finishedSqls.length > 0) {
           // in cases of SQLs out of order, like id 2 is running and 3 is completed, we will try to ask from id 2 again
           finishedSqls.forEach((sql) => {
-            if (parseInt(sql.id) === this.lastCompletedSqlId + 1) {
+            const idAsNumber = parseInt(sql.id)
+            if (idAsNumber === this.lastCompletedSqlId + 1) {
               this.lastCompletedSqlId += 1;
             }
+            if (this.sqlIdToFinishTime[idAsNumber] === undefined) {
+              this.sqlIdToFinishTime[idAsNumber] = { id: idAsNumber, finishTime: timeStrToEpocTime(sql.submissionTime) + sql.duration };
+            }
           });
         }