tests: separate performance tests

Before this patch, performance tests ran together with unit and
integration with `--coverage` flag. Coverage analysis cropped the
result of performance tests to 10-15 times. For metrics integration
it resulted in timeout errors and drop of performance which is not
reproduces with coverage disabled. Moreover, before this patch log
capture was disabled and performance tests did not displayed any
results after run. Now performance tests also run is separate CI job.

After this patch, `make -C build coverage` will run lightweight
version of performance test. `make -C build performance` will run real
performance tests.

You can paste output table to GitHub [1].

This path also reworks current performance test. It adds new cases to
compare module performance with or without statistics, statistic
wrappers and compare different metrics drivers and reports new info:
average call time and max call time.

Performance test result: overhead is 3-10% in case of `local` driver and
5-15% in case of `metrics` driver, up to 20% for `metrics` with
quantiles. Based on several runs on HP ProBook 440 G7 i7/16Gb/256SSD.

1. https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/organizing-information-with-tables

Closes #233, follows up #224

.github/workflows/test_on_push.yaml

@@ -78,6 +78,41 @@ jobs:
         run: make -C build coveralls
         if: ${{ matrix.coveralls }}
 
+  run-perf-tests-ce:
+    if: |
+      github.event_name == 'push' ||
+      github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name != github.repository
+    strategy:
+      matrix:
+        tarantool-version: ["1.10", "2.8"]
+        metrics-version: ["0.12.0"]
+      fail-fast: false
+    runs-on: [ubuntu-latest]
+    steps:
+      - uses: actions/checkout@master
+
+      - name: Setup Tarantool CE 
+        uses: tarantool/setup-tarantool@v1
+        with:
+          tarantool-version: ${{ matrix.tarantool-version }}
+
+      - name: Install requirements for community
+        run: |
+          tarantool --version
+          ./deps.sh
+
+      - name: Install metrics
+        run: tarantoolctl rocks install metrics ${{ matrix.metrics-version }}
+
+      # This server starts and listen on 8084 port that is used for tests
+      - name: Stop Mono server
+        run: sudo kill -9 $(sudo lsof -t -i tcp:8084) || true
+
+      - run: cmake -S . -B build
+
+      - name: Run performance tests
+        run: make -C build performance
+
   run-tests-ee:
     if: github.event_name == 'push'
     strategy:

CMakeLists.txt

@@ -36,6 +36,14 @@ add_custom_target(luatest
   COMMENT "Run regression tests"
 )
 
+set(PERFORMANCE_TESTS_SUBDIR "test/performance")
+
+add_custom_target(performance
+  COMMAND PERF_MODE_ON=true ${LUATEST} -v -c ${PERFORMANCE_TESTS_SUBDIR}
+  WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}
+  COMMENT "Run performance tests"
+)
+
 add_custom_target(coverage
   COMMAND ${LUACOV} ${PROJECT_SOURCE_DIR} && grep -A999 '^Summary' ${CODE_COVERAGE_REPORT}
   DEPENDS ${CODE_COVERAGE_STATS}

README.md

@@ -718,6 +718,9 @@ crud.cfg
   stats_driver: local
 ...
 ```
+Performance overhead is 3-10% in case of `local` driver and
+5-15% in case of `metrics` driver, up to 20% for `metrics` with quantiles.
+
 Beware that iterating through `crud.cfg` with pairs is not supported yet,
 refer to [tarantool/crud#265](https://github.com/tarantool/crud/issues/265).
 

test/entrypoint/srv_ddl.lua

@@ -102,6 +102,12 @@ package.preload['customers-storage'] = function()
                 },
             }
 
+            local customers_id_schema = table.deepcopy(customers_schema)
+            customers_id_schema.sharding_key = {'id'}
+            table.insert(customers_id_schema.indexes, primary_index_id)
+            table.insert(customers_id_schema.indexes, bucket_id_index)
+            table.insert(customers_id_schema.indexes, age_index)
+
             local customers_name_key_schema = table.deepcopy(customers_schema)
             customers_name_key_schema.sharding_key = {'name'}
             table.insert(customers_name_key_schema.indexes, primary_index)
@@ -157,6 +163,7 @@ package.preload['customers-storage'] = function()
 
             local schema = {
                 spaces = {
+                    customers = customers_id_schema,
                     customers_name_key = customers_name_key_schema,
                     customers_name_key_uniq_index = customers_name_key_uniq_index_schema,
                     customers_name_key_non_uniq_index = customers_name_key_non_uniq_index_schema,
@@ -195,8 +202,13 @@ local ok, err = errors.pcall('CartridgeCfgError', cartridge.cfg, {
         'customers-storage',
         'cartridge.roles.crud-router',
         'cartridge.roles.crud-storage',
-    },
-})
+    }},
+    -- Increase readahead for performance tests.
+    -- Performance tests on HP ProBook 440 G5 16 Gb
+    -- bump into default readahead limit and thus not
+    -- give a full picture.
+    { readahead = 20 * 1024 * 1024 }
+)
 
 if not ok then
     log.error('%s', err)

test/integration/ddl_sharding_key_test.lua

@@ -696,6 +696,7 @@ pgroup.test_update_cache_with_incorrect_key = function(g)
     -- records for all spaces exist
     sharding_key_as_index_obj = helpers.get_sharding_key_cache(g.cluster)
     t.assert_equals(sharding_key_as_index_obj, {
+        customers = {parts = {{fieldno = 1}}},
         customers_G_func = {parts = {{fieldno = 1}}},
         customers_body_func = {parts = {{fieldno = 1}}},
         customers_age_key = {parts = {{fieldno = 4}}},
@@ -722,6 +723,7 @@ pgroup.test_update_cache_with_incorrect_key = function(g)
     -- other records for correct spaces exist in cache
     sharding_key_as_index_obj = helpers.get_sharding_key_cache(g.cluster)
     t.assert_equals(sharding_key_as_index_obj, {
+        customers = {parts = {{fieldno = 1}}},
         customers_G_func = {parts = {{fieldno = 1}}},
         customers_body_func = {parts = {{fieldno = 1}}},
         customers_age_key = {parts = {{fieldno = 4}}},
@@ -747,6 +749,7 @@ pgroup.test_update_cache_with_incorrect_key = function(g)
     -- other records for correct spaces exist in cache
     sharding_key_as_index_obj = helpers.get_sharding_key_cache(g.cluster)
     t.assert_equals(sharding_key_as_index_obj, {
+        customers = {parts = {{fieldno = 1}}},
         customers_G_func = {parts = {{fieldno = 1}}},
         customers_body_func = {parts = {{fieldno = 1}}},
         customers_age_key = {parts = {{fieldno = 4}}},