diff --git a/.github/ISSUE_TEMPLATE/question.md b/.github/ISSUE_TEMPLATE/general-question.md
similarity index 78%
rename from .github/ISSUE_TEMPLATE/question.md
rename to .github/ISSUE_TEMPLATE/general-question.md
index ec486cc277fdd..22c9478417525 100644
--- a/.github/ISSUE_TEMPLATE/question.md
+++ b/.github/ISSUE_TEMPLATE/general-question.md
@@ -1,10 +1,10 @@
 ---
-name: "\U0001F914 Question"
+name: "\U0001F914 General Question"
 about: Usage question that isn't answered in docs or discussion
 
 ---
 
-## Question
+## General Question
 
 Before asking a question, make sure you have:
 
@@ -13,4 +13,5 @@ Before asking a question, make sure you have:
 - Searched open and closed [GitHub issues](https://github.com/pingcap/tidb/issues?utf8=%E2%9C%93&q=is%3Aissue)
 - Read the documentation:
   - [TiDB Readme](https://github.com/pingcap/tidb)
-  - [TiDB Doc](https://github.com/pingcap/docs)
\ No newline at end of file
+  - [TiDB Doc](https://github.com/pingcap/docs)
+
diff --git a/.github/ISSUE_TEMPLATE/performance-questions.md b/.github/ISSUE_TEMPLATE/performance-questions.md
new file mode 100644
index 0000000000000..053386ac55c21
--- /dev/null
+++ b/.github/ISSUE_TEMPLATE/performance-questions.md
@@ -0,0 +1,28 @@
+---
+name: "\U0001F947 Performance Questions"
+about: Performance question about TiDB which is not caused by bug.
+---
+
+## Performance Questions
+
+- What version of TiDB are you using?
+<!-- You can try `tidb-server -V` or run `select tidb_version();` on TiDB to get this information -->
+
+- What's the observed and your expected performance respectively?
+
+- Have you compared TiDB with other databases? If yes, what's their difference?
+
+- For a specific slow SQL query, please provide the following information:
+    - Whether you analyzed the tables involved in the query and how long it is after you ran the last `ANALYZE`.
+    - Whether this SQL query always or occasionally runs slowly.
+    - The `EXPLAIN ANALYZE` result of this query if your TiDB version is higher than 2.1, or you can just provide the `EXPLAIN` result.
+    - The plain text of the SQL query and table schema so we can test it locally. It would be better if you can provide the dumped statistics information.
+        <!-- you can use `show create table ${involved_table}\G` to get the table schema.-->
+        <!-- use `curl -G "http://${tidb-server-ip}:${tidb-server-status-port}/stats/dump/${db_name}/${table_name}" > ${table_name}_stats.json` to get the dumped statistics of one involved table.-->
+    - The `EXPLAIN` result of the compared database. For MySQL, `EXPLAIN format=json`'s result will be more helpful.
+    - Other information that is useful from your perspective.
+
+- For a general performance question, e.g. the benchmark result you got by yourself is not expected, please provide the following information:
+    - Your cluster's topology architecture.
+    - A simple description of you workload.
+    - The metrics PDF generated from Grafana monitor. Remember to set the time range to the performance issue duration.