Merge branch 'master' into enh/search-space

Author: odeimaiz

.env-devel

@@ -50,6 +50,7 @@ CLUSTERS_KEEPER_COMPUTATIONAL_BACKEND_DOCKER_IMAGE_TAG=master-github-latest
 CLUSTERS_KEEPER_DASK_NTHREADS=0
 CLUSTERS_KEEPER_DASK_WORKER_SATURATION=inf
 CLUSTERS_KEEPER_EC2_ACCESS=null
+CLUSTERS_KEEPER_SSM_ACCESS=null
 CLUSTERS_KEEPER_EC2_INSTANCES_PREFIX=""
 CLUSTERS_KEEPER_LOGLEVEL=WARNING
 CLUSTERS_KEEPER_MAX_MISSED_HEARTBEATS_BEFORE_CLUSTER_TERMINATION=5

.github/workflows/ci-testing-deploy.yml

@@ -590,7 +590,7 @@ jobs:
   unit-test-autoscaling:
     needs: changes
     if: ${{ needs.changes.outputs.autoscaling == 'true' || github.event_name == 'push' }}
-    timeout-minutes: 19 # if this timeout gets too small, then split the tests
+    timeout-minutes: 22 # temporary: mypy takes a huge amount of time to run here, maybe we should cache it
     name: "[unit] autoscaling"
     runs-on: ${{ matrix.os }}
     strategy:

packages/models-library/src/models_library/clusters.py

@@ -96,6 +96,9 @@ class Config(BaseAuthentication.Config):
 class NoAuthentication(BaseAuthentication):
     type: Literal["none"] = "none"
 
+    class Config(BaseAuthentication.Config):
+        schema_extra: ClassVar[dict[str, Any]] = {"examples": [{"type": "none"}]}
+
 
 class TLSAuthentication(BaseAuthentication):
     type: Literal["tls"] = "tls"

packages/models-library/src/models_library/users.py

@@ -22,7 +22,7 @@ class UserBillingDetails(BaseModel):
     address: str | None
     city: str | None
     state: str | None = Field(description="State, province, canton, ...")
-    country: str
+    country: str  # Required for taxes
     postal_code: str | None
     phone: str | None
 

packages/pytest-simcore/src/pytest_simcore/helpers/logging_tools.py

@@ -133,14 +133,14 @@ def log_context(
     else:
         ctx_msg = msg
 
-    started_time = datetime.datetime.now(tz=datetime.timezone.utc)
+    started_time = datetime.datetime.now(tz=datetime.UTC)
     try:
         DynamicIndentFormatter.cls_increase_indent()
 
         logger.log(level, ctx_msg.starting, *args, **kwargs)
         with _increased_logger_indent(logger):
             yield SimpleNamespace(logger=logger, messages=ctx_msg)
-        elapsed_time = datetime.datetime.now(tz=datetime.timezone.utc) - started_time
+        elapsed_time = datetime.datetime.now(tz=datetime.UTC) - started_time
         done_message = (
             f"{ctx_msg.done} ({_timedelta_as_minute_second_ms(elapsed_time)})"
         )
@@ -152,7 +152,7 @@ def log_context(
         )
 
     except:
-        elapsed_time = datetime.datetime.now(tz=datetime.timezone.utc) - started_time
+        elapsed_time = datetime.datetime.now(tz=datetime.UTC) - started_time
         error_message = (
             f"{ctx_msg.raised} ({_timedelta_as_minute_second_ms(elapsed_time)})"
         )

packages/pytest-simcore/src/pytest_simcore/helpers/playwright.py

@@ -3,10 +3,10 @@
 import logging
 import re
 from collections import defaultdict
-from contextlib import ExitStack
+from collections.abc import Generator, Iterator
 from dataclasses import dataclass, field
 from enum import Enum, unique
-from typing import Any, Final, Generator
+from typing import Any, Final
 
 from playwright.sync_api import FrameLocator, Page, Request, WebSocket
 from pytest_simcore.helpers.logging_tools import log_context
@@ -263,28 +263,37 @@ def wait_for_pipeline_state(
     return current_state
 
 
-def on_web_socket_default_handler(ws) -> None:
-    """Usage
-
-    from pytest_simcore.playwright_utils import on_web_socket_default_handler
-
-    page.on("websocket", on_web_socket_default_handler)
-
-    """
-    stack = ExitStack()
-    ctx = stack.enter_context(
-        log_context(
-            logging.INFO,
-            (
-                f"WebSocket opened: {ws.url}",
-                "WebSocket closed",
-            ),
-        )
-    )
+@contextlib.contextmanager
+def web_socket_default_log_handler(web_socket: WebSocket) -> Iterator[None]:
 
-    ws.on("framesent", lambda payload: ctx.logger.info("⬇️ %s", payload))
-    ws.on("framereceived", lambda payload: ctx.logger.info("⬆️ %s", payload))
-    ws.on("close", lambda payload: stack.close())  # noqa: ARG005
+    try:
+        with log_context(
+            logging.DEBUG,
+            msg="handle websocket message (set to --log-cli-level=DEBUG level if you wanna see all of them)",
+        ) as ctx:
+
+            def on_framesent(payload: str | bytes) -> None:
+                ctx.logger.debug("⬇️ Frame sent: %s", payload)
+
+            def on_framereceived(payload: str | bytes) -> None:
+                ctx.logger.debug("⬆️ Frame received: %s", payload)
+
+            def on_close(payload: WebSocket) -> None:
+                ctx.logger.warning("⚠️ Websocket closed: %s", payload)
+
+            def on_socketerror(error_msg: str) -> None:
+                ctx.logger.error("❌ Websocket error: %s", error_msg)
+
+            web_socket.on("framesent", on_framesent)
+            web_socket.on("framereceived", on_framereceived)
+            web_socket.on("close", on_close)
+            web_socket.on("socketerror", on_socketerror)
+            yield
+    finally:
+        web_socket.remove_listener("framesent", on_framesent)
+        web_socket.remove_listener("framereceived", on_framereceived)
+        web_socket.remove_listener("close", on_close)
+        web_socket.remove_listener("socketerror", on_socketerror)
 
 
 def _node_started_predicate(request: Request) -> bool:

packages/pytest-simcore/src/pytest_simcore/helpers/playwright_sim4life.py

@@ -6,6 +6,8 @@
 
 import arrow
 from playwright.sync_api import FrameLocator, Page, WebSocket, expect
+from pydantic import TypeAdapter  # pylint: disable=no-name-in-module
+from pydantic import ByteSize
 
 from .logging_tools import log_context
 from .playwright import (
@@ -17,7 +19,7 @@
     wait_for_service_running,
 )
 
-_S4L_STREAMING_ESTABLISHMENT_MAX_TIME: Final[int] = 15 * SECOND
+_S4L_STREAMING_ESTABLISHMENT_MAX_TIME: Final[int] = 30 * SECOND
 _S4L_SOCKETIO_REGEX: Final[re.Pattern] = re.compile(
     r"^(?P<protocol>[^:]+)://(?P<node_id>[^\.]+)\.services\.(?P<hostname>[^\/]+)\/socket\.io\/.+$"
 )
@@ -63,7 +65,7 @@ def __call__(self, message: str) -> bool:
                     self._initial_bit_rate_time = arrow.utcnow().datetime
                     self.logger.info(
                         "%s",
-                        f"{self._initial_bit_rate=} at {self._initial_bit_rate_time.isoformat()}",
+                        f"{TypeAdapter(ByteSize).validate_python(self._initial_bit_rate).human_readable()}/s at {self._initial_bit_rate_time.isoformat()}",
                     )
                     return False
 
@@ -78,7 +80,7 @@ def __call__(self, message: str) -> bool:
                     bitrate_test = bool(self._initial_bit_rate != current_bitrate)
                     self.logger.info(
                         "%s",
-                        f"{current_bitrate=} after {elapsed_time=}: {'good!' if bitrate_test else 'failed! bitrate did not change! TIP: talk with MaG about underwater cables!'}",
+                        f"{TypeAdapter(ByteSize).validate_python(current_bitrate).human_readable()}/s after {elapsed_time=}: {'good!' if bitrate_test else 'failed! bitrate did not change! TIP: talk with MaG about underwater cables!'}",
                     )
                     return bitrate_test
 

packages/service-library/src/servicelib/fastapi/prometheus_instrumentation.py

@@ -2,23 +2,27 @@
 
 
 from fastapi import FastAPI
+from prometheus_client import CollectorRegistry
 from prometheus_fastapi_instrumentator import Instrumentator
 
 
 def setup_prometheus_instrumentation(app: FastAPI) -> Instrumentator:
+    # NOTE: use that registry to prevent having a global one
+    app.state.prometheus_registry = registry = CollectorRegistry(auto_describe=True)
+    instrumentator = Instrumentator(
+        should_instrument_requests_inprogress=False,  # bug in https://github.com/trallnag/prometheus-fastapi-instrumentator/issues/317
+        inprogress_labels=False,
+        registry=registry,
+    ).instrument(app)
 
-    instrumentator = (
-        Instrumentator(
-            should_instrument_requests_inprogress=True, inprogress_labels=False
-        )
-        .instrument(app)
-        .expose(app, include_in_schema=False)
-    )
+    async def _on_startup() -> None:
+        instrumentator.expose(app, include_in_schema=False)
 
-    def _unregister():
-        for collector in list(instrumentator.registry._collector_to_names.keys()):
-            instrumentator.registry.unregister(collector)
+    def _unregister() -> None:
+        # NOTE: avoid registering collectors multiple times when running unittests consecutively (https://stackoverflow.com/a/62489287)
+        for collector in list(registry._collector_to_names.keys()):  # noqa: SLF001
+            registry.unregister(collector)
 
-    # avoid registering collectors multiple times when running unittests consecutively (https://stackoverflow.com/a/62489287)
+    app.add_event_handler("startup", _on_startup)
     app.add_event_handler("shutdown", _unregister)
     return instrumentator

packages/service-library/src/servicelib/fastapi/tracing.py

@@ -1,16 +1,15 @@
 """ Adds fastapi middleware for tracing using opentelemetry instrumentation.
 
 """
+
 import logging
 
 from fastapi import FastAPI
 from opentelemetry import trace
 from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
     OTLPSpanExporter as OTLPSpanExporterHTTP,
 )
-from opentelemetry.instrumentation.fastapi import (
-    FastAPIInstrumentor,  # pylint: disable=no-name-in-module
-)
+from opentelemetry.instrumentation.fastapi import FastAPIInstrumentor
 from opentelemetry.sdk.resources import Resource
 from opentelemetry.sdk.trace import TracerProvider
 from opentelemetry.sdk.trace.export import BatchSpanProcessor
@@ -21,24 +20,19 @@
 
 def setup_tracing(
     app: FastAPI, tracing_settings: TracingSettings, service_name: str
-) -> FastAPIInstrumentor | None:
+) -> None:
     if (
         not tracing_settings.TRACING_OPENTELEMETRY_COLLECTOR_ENDPOINT
         and not tracing_settings.TRACING_OPENTELEMETRY_COLLECTOR_PORT
     ):
         log.warning("Skipping opentelemetry tracing setup")
-        return None
-    if (
-        not tracing_settings.TRACING_OPENTELEMETRY_COLLECTOR_ENDPOINT
-        or not tracing_settings.TRACING_OPENTELEMETRY_COLLECTOR_PORT
-    ):
-        raise RuntimeError(
-            f"Variable opentelemetry_collector_endpoint [{tracing_settings.TRACING_OPENTELEMETRY_COLLECTOR_ENDPOINT}] or opentelemetry_collector_port [{tracing_settings.TRACING_OPENTELEMETRY_COLLECTOR_PORT}] unset. Tracing options incomplete."
-        )
+        return
+
     # Set up the tracer provider
     resource = Resource(attributes={"service.name": service_name})
     trace.set_tracer_provider(TracerProvider(resource=resource))
-    tracer_provider = trace.get_tracer_provider()
+    global_tracer_provider = trace.get_tracer_provider()
+    assert isinstance(global_tracer_provider, TracerProvider)  # nosec
     tracing_destination: str = f"{tracing_settings.TRACING_OPENTELEMETRY_COLLECTOR_ENDPOINT}:{tracing_settings.TRACING_OPENTELEMETRY_COLLECTOR_PORT}/v1/traces"
     log.info(
         "Trying to connect service %s to tracing collector at %s.",
@@ -48,7 +42,6 @@ def setup_tracing(
     # Configure OTLP exporter to send spans to the collector
     otlp_exporter = OTLPSpanExporterHTTP(endpoint=tracing_destination)
     span_processor = BatchSpanProcessor(otlp_exporter)
-    # Mypy bug --> https://github.com/open-telemetry/opentelemetry-python/issues/3713
-    tracer_provider.add_span_processor(span_processor)  # type: ignore[attr-defined]
+    global_tracer_provider.add_span_processor(span_processor)
     # Instrument FastAPI
-    return FastAPIInstrumentor().instrument_app(app)  # type: ignore[no-any-return]
+    FastAPIInstrumentor().instrument_app(app)

packages/service-library/src/servicelib/instrumentation.py

@@ -1,2 +1,13 @@
+from dataclasses import dataclass
+
+from prometheus_client import CollectorRegistry
+
+
+@dataclass(slots=True, kw_only=True)
+class MetricsBase:
+    subsystem: str
+    registry: CollectorRegistry
+
+
 def get_metrics_namespace(application_name: str) -> str:
     return application_name.replace("-", "_")