WIP: Query Service SLO

.github/workflows/slo.yml

@@ -31,6 +31,7 @@ jobs:
         if: env.DOCKER_REPO != null
         env:
           DOCKER_REPO: ${{ secrets.SLO_DOCKER_REPO }}
+        continue-on-error: true
         with:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           KUBECONFIG_B64: ${{ secrets.SLO_KUBE_CONFIG }}
@@ -48,11 +49,17 @@ jobs:
           timeBetweenPhases: 30
           shutdownTime: 30
 
-          language_id0: sync
-          language0: python-sync
+          language_id0: sync-python-table
+          language0: Python SDK over Table Service
           workload_path0: tests/slo
           workload_build_context0: ../..
-          workload_build_options0: -f Dockerfile
+          workload_build_options0: -f Dockerfile --build-arg SDK_SERVICE=sync-python-table
+
+          language_id1: sync-python-query
+          language1: Python SDK over Query Service
+          workload_path1: tests/slo
+          workload_build_context1: ../..
+          workload_build_options1: -f Dockerfile --build-arg SDK_SERVICE=sync-python-query
 
       - uses: actions/upload-artifact@v3
         if: env.DOCKER_REPO != null

tests/slo/Dockerfile

@@ -3,5 +3,7 @@ COPY . /src
 WORKDIR /src
 RUN python -m pip install --upgrade pip && python -m pip install -e . && python -m pip install -r tests/slo/requirements.txt
 WORKDIR tests/slo
+ARG SDK_SERVICE
+ENV SDK_SERVICE=$SDK_SERVICE
 
 ENTRYPOINT ["python", "src"]

tests/slo/src/jobs.py

@@ -3,7 +3,7 @@
 import logging
 import dataclasses
 from random import randint
-from typing import Callable, Tuple
+from typing import Callable, Tuple, Union
 from ratelimiter import RateLimiter
 
 import threading
@@ -31,12 +31,24 @@
 );
 """
 
+# QUERY_READ_QUERY_TEMPLATE = """
+# SELECT * FROM `{}` WHERE object_id = $object_id AND object_hash = Digest::NumericHash($object_id);
+# """
+
+# QUERY_WRITE_QUERY_TEMPLATE = """
+# UPSERT INTO `{}` (
+#     object_id, object_hash, payload_str, payload_double, payload_timestamp
+# ) VALUES (
+#     $object_id, Digest::NumericHash($object_id), $payload_str, $payload_double, $payload_timestamp
+# );
+# """
+
 logger = logging.getLogger(__name__)
 
 
 @dataclasses.dataclass
 class RequestParams:
-    pool: ydb.SessionPool
+    pool: Union[ydb.SessionPool, ydb.QuerySessionPool]
     query: str
     params: dict
     metrics: Metrics
@@ -56,7 +68,7 @@ def transaction(session):
 
         result = session.transaction().execute(
             params.query,
-            params.params,
+            parameters=params.params,
             commit_tx=True,
             settings=params.request_settings,
         )
@@ -82,7 +94,7 @@ def transaction(session):
 def run_reads(driver, query, max_id, metrics, limiter, runtime, timeout):
     start_time = time.time()
 
-    logger.info("Start read workload")
+    logger.info("Start read workload over table service")
 
     request_settings = ydb.BaseRequestSettings().with_timeout(timeout)
     retry_setting = ydb.RetrySettings(
@@ -116,7 +128,7 @@ def check_result(result):
 
 
 def run_read_jobs(args, driver, tb_name, max_id, metrics):
-    logger.info("Start read jobs")
+    logger.info("Start read jobs over table service")
 
     session = ydb.retry_operation_sync(lambda: driver.table_client.session().create())
     read_q = session.prepare(READ_QUERY_TEMPLATE.format(tb_name))
@@ -135,10 +147,65 @@ def run_read_jobs(args, driver, tb_name, max_id, metrics):
     return futures
 
 
+def run_reads_query(driver, query, max_id, metrics, limiter, runtime, timeout):
+    start_time = time.time()
+
+    logger.info("Start read workload over query service")
+
+    request_settings = ydb.BaseRequestSettings().with_timeout(timeout)
+    retry_setting = ydb.RetrySettings(
+        idempotent=True,
+        max_session_acquire_timeout=timeout,
+    )
+
+    with ydb.QuerySessionPool(driver) as pool:
+        logger.info("Session pool for read requests created")
+
+        while time.time() - start_time < runtime:
+            params = {"$object_id": (randint(1, max_id), ydb.PrimitiveType.Uint64)}
+            with limiter:
+
+                def check_result(result):
+                    res = next(result)
+                    assert res.rows[0]
+
+                params = RequestParams(
+                    pool=pool,
+                    query=query,
+                    params=params,
+                    metrics=metrics,
+                    labels=(JOB_READ_LABEL,),
+                    request_settings=request_settings,
+                    retry_settings=retry_setting,
+                    check_result_cb=check_result,
+                )
+                execute_query(params)
+
+        logger.info("Stop read workload")
+
+
+def run_read_jobs_query(args, driver, tb_name, max_id, metrics):
+    logger.info("Start read jobs over query service")
+
+    read_q = READ_QUERY_TEMPLATE.format(tb_name)
+
+    read_limiter = RateLimiter(max_calls=args.read_rps, period=1)
+    futures = []
+    for _ in range(args.read_threads):
+        future = threading.Thread(
+            name="slo_run_read",
+            target=run_reads_query,
+            args=(driver, read_q, max_id, metrics, read_limiter, args.time, args.read_timeout / 1000),
+        )
+        future.start()
+        futures.append(future)
+    return futures
+
+
 def run_writes(driver, query, row_generator, metrics, limiter, runtime, timeout):
     start_time = time.time()
 
-    logger.info("Start write workload")
+    logger.info("Start write workload over table service")
 
     request_settings = ydb.BaseRequestSettings().with_timeout(timeout)
     retry_setting = ydb.RetrySettings(
@@ -157,6 +224,7 @@ def run_writes(driver, query, row_generator, metrics, limiter, runtime, timeout)
                 "$payload_double": row.payload_double,
                 "$payload_timestamp": row.payload_timestamp,
             }
+
             with limiter:
                 params = RequestParams(
                     pool=pool,
@@ -173,7 +241,7 @@ def run_writes(driver, query, row_generator, metrics, limiter, runtime, timeout)
 
 
 def run_write_jobs(args, driver, tb_name, max_id, metrics):
-    logger.info("Start write jobs")
+    logger.info("Start write jobs over table service")
 
     session = ydb.retry_operation_sync(lambda: driver.table_client.session().create())
     write_q = session.prepare(WRITE_QUERY_TEMPLATE.format(tb_name))
@@ -194,6 +262,69 @@ def run_write_jobs(args, driver, tb_name, max_id, metrics):
     return futures
 
 
+def run_writes_query(driver, query, row_generator, metrics, limiter, runtime, timeout):
+    start_time = time.time()
+
+    logger.info("Start write workload over query service")
+
+    request_settings = ydb.BaseRequestSettings().with_timeout(timeout)
+    retry_setting = ydb.RetrySettings(
+        idempotent=True,
+        max_session_acquire_timeout=timeout,
+    )
+
+    with ydb.QuerySessionPool(driver) as pool:
+        logger.info("Session pool for read requests created")
+
+        while time.time() - start_time < runtime:
+            row = row_generator.get()
+            params = {
+                "$object_id": (row.object_id, ydb.PrimitiveType.Uint64),
+                "$payload_str": (row.payload_str, ydb.PrimitiveType.Utf8),
+                "$payload_double": (row.payload_double, ydb.PrimitiveType.Double),
+                "$payload_timestamp": (row.payload_timestamp, ydb.PrimitiveType.Timestamp),
+            }
+
+            def check_result(result):
+                with result:
+                    pass
+
+            with limiter:
+                params = RequestParams(
+                    pool=pool,
+                    query=query,
+                    params=params,
+                    metrics=metrics,
+                    labels=(JOB_WRITE_LABEL,),
+                    request_settings=request_settings,
+                    retry_settings=retry_setting,
+                    check_result_cb=check_result,
+                )
+                execute_query(params)
+
+        logger.info("Stop write workload")
+
+
+def run_write_jobs_query(args, driver, tb_name, max_id, metrics):
+    logger.info("Start write jobs for query service")
+
+    write_q = WRITE_QUERY_TEMPLATE.format(tb_name)
+
+    write_limiter = RateLimiter(max_calls=args.write_rps, period=1)
+    row_generator = RowGenerator(max_id)
+
+    futures = []
+    for _ in range(args.write_threads):
+        future = threading.Thread(
+            name="slo_run_write",
+            target=run_writes_query,
+            args=(driver, write_q, row_generator, metrics, write_limiter, args.time, args.write_timeout / 1000),
+        )
+        future.start()
+        futures.append(future)
+    return futures
+
+
 def push_metric(limiter, runtime, metrics):
     start_time = time.time()
     logger.info("Start push metrics")

tests/slo/src/metrics.py

@@ -13,6 +13,9 @@
 JOB_READ_LABEL, JOB_WRITE_LABEL = "read", "write"
 JOB_STATUS_OK, JOB_STATUS_ERR = "ok", "err"
 
+SDK_SERVICE_NAME = environ.get("SDK_SERVICE", "sync-python-table")
+print(f"SDK_SERVICE_NAME = {SDK_SERVICE_NAME}")
+
 
 class Metrics:
     def __init__(self, push_gateway):
@@ -102,10 +105,10 @@ def stop(self, labels, start_time, attempts=1, error=None):
     def push(self):
         push_to_gateway(
             self._push_gtw,
-            job="workload-sync",
+            job=f"workload-{SDK_SERVICE_NAME}",
             registry=self._registry,
             grouping_key={
-                "sdk": "python-sync",
+                "sdk": SDK_SERVICE_NAME,
                 "sdkVersion": version("ydb"),
             },
         )

tests/slo/src/runner.py

@@ -7,8 +7,14 @@
 import concurrent.futures
 from concurrent.futures import ThreadPoolExecutor
 
-from jobs import run_read_jobs, run_write_jobs, run_metric_job
-from metrics import Metrics
+from jobs import (
+    run_read_jobs,
+    run_write_jobs,
+    run_read_jobs_query,
+    run_write_jobs_query,
+    run_metric_job,
+)
+from metrics import Metrics, SDK_SERVICE_NAME
 
 logger = logging.getLogger(__name__)
 
@@ -85,17 +91,25 @@ def run_slo(args, driver, tb_name):
     logger.info("Max ID: %s", max_id)
 
     metrics = Metrics(args.prom_pgw)
-
-    futures = (
-        *run_read_jobs(args, driver, tb_name, max_id, metrics),
-        *run_write_jobs(args, driver, tb_name, max_id, metrics),
-        run_metric_job(args, metrics),
-    )
+    if SDK_SERVICE_NAME == "sync-python-table":
+        futures = (
+            *run_read_jobs(args, driver, tb_name, max_id, metrics),
+            *run_write_jobs(args, driver, tb_name, max_id, metrics),
+            run_metric_job(args, metrics),
+        )
+    elif SDK_SERVICE_NAME == "sync-python-query":
+        futures = (
+            *run_read_jobs_query(args, driver, tb_name, max_id, metrics),
+            *run_write_jobs_query(args, driver, tb_name, max_id, metrics),
+            run_metric_job(args, metrics),
+        )
+    else:
+        raise ValueError(f"Unsupported service: {SDK_SERVICE_NAME}")
 
     for future in futures:
         future.join()
 
-    metrics.reset()
+    # metrics.reset()
 
 
 def run_cleanup(args, driver, tb_name):
@@ -114,7 +128,7 @@ def run_from_args(args):
     table_name = path.join(args.db, args.table_name)
 
     with ydb.Driver(driver_config) as driver:
-        driver.wait(timeout=5)
+        driver.wait(timeout=300)
         try:
             if args.subcommand == "create":
                 run_create(args, driver, table_name)