Send anonymous telemetry to bugout (pantsbuild#11697)

Uses the `humbug` library to send anonymized telemetry data
to bugout.

Repos must opt in to telemetry. We log a warning if a repo has
not opted in or out.

See the code and also https://www.pantsbuild.org/v2.4/docs/anonymous-telemetry
for more info on the data we send and how we preserve anonymity
and prevent leakage of proprietary information.

[ci skip-rust]

[ci skip-build-wheels]

3rdparty/python/constraints.txt

@@ -1,25 +1,28 @@
-# Generated by build-support/bin/generate_lockfile.sh on Tue Mar 16 06:41:23 PM PDT 2021
+# Generated by build-support/bin/generate_lockfile.sh on Wed Mar 17 11:10:48 PDT 2021
 ansicolors==1.1.8
 attrs==20.3.0
 beautifulsoup4==4.6.3
+bugout==0.1.8
 certifi==2020.12.5
 cffi==1.14.5
 chardet==4.0.0
 cryptography==3.4.6
 fasteners==0.15
 freezegun==1.0.0
+humbug==0.1.9
 idna==2.10
 iniconfig==1.1.1
 monotonic==1.5
 mypy==0.800
 mypy-extensions==0.4.3
 packaging==20.9
 pex==2.1.34
-pip==20.2.3
+pip==20.2.4
 pluggy==0.13.1
 psutil==5.7.0
 py==1.10.0
 pycparser==2.20
+pydantic==1.8.1
 pyOpenSSL==20.0.1
 pyparsing==2.4.7
 pystache==0.5.4
@@ -32,5 +35,5 @@ setuptools==53.1.0
 six==1.15.0
 toml==0.10.2
 typed-ast==1.4.2
-typing-extensions==3.7.4.2
+typing-extensions==3.7.4.3
 urllib3==1.26.4

3rdparty/python/requirements.txt

@@ -3,6 +3,10 @@ beautifulsoup4>=4.6.0,<4.7
 fasteners==0.15.0
 freezegun==1.0.0
 
+# Note: we use humbug to report telemetry. When upgrading, ensure the new version maintains the
+# anonymity promise we make here: https://www.pantsbuild.org/docs/anonymous-telemetry
+humbug==0.1.9
+
 # The MyPy requirement should be maintained in lockstep with the requirement the Pants repo uses
 # for the mypy task since it configures custom MyPy plugins. That requirement can be found via:
 #   ./pants help-all | \
@@ -25,4 +29,4 @@ requests[security]>=2.20.1
 setproctitle==1.2
 setuptools>=50.3.0,<54.0
 toml==0.10.2
-typing-extensions==3.7.4.2
+typing-extensions==3.7.4.3

pants.toml

@@ -53,6 +53,10 @@ pants_ignore.add = [
   "!/pants.pex",
 ]
 
+[anonymous-telemetry]
+enabled = true
+repo_id = "7775F8D5-FC58-4DBC-9302-D00AE4A1505F"
+
 [source]
 root_patterns = [
   "src/*",

src/python/pants/core/register.py

@@ -19,7 +19,7 @@
     stripped_source_files,
     subprocess_environment,
 )
-from pants.goal import stats_aggregator
+from pants.goal import anonymous_telemetry, stats_aggregator
 from pants.source import source_root
 
 
@@ -45,6 +45,7 @@ def rules():
         *subprocess_environment.rules(),
         *source_root.rules(),
         *target_type_rules(),
+        *anonymous_telemetry.rules(),
         *stats_aggregator.rules(),
     ]
 

src/python/pants/goal/anonymous_telemetry.py

@@ -0,0 +1,192 @@
+# Copyright 2021 Pants project contributors (see CONTRIBUTORS.md).
+# Licensed under the Apache License, Version 2.0 (see LICENSE).
+
+from __future__ import annotations
+
+import json
+import logging
+import re
+import uuid
+from typing import cast
+
+from humbug.consent import HumbugConsent  # type: ignore
+from humbug.report import Modes, Report, Reporter  # type: ignore
+
+from pants.engine.internals.scheduler import Workunit
+from pants.engine.rules import collect_rules, rule
+from pants.engine.streaming_workunit_handler import (
+    StreamingWorkunitContext,
+    WorkunitsCallback,
+    WorkunitsCallbackFactory,
+    WorkunitsCallbackFactoryRequest,
+)
+from pants.engine.unions import UnionRule
+from pants.option.subsystem import Subsystem
+from pants.util.docutil import docs_url
+
+logger = logging.getLogger(__name__)
+
+
+_bugout_access_token = "3ae76900-9a68-4a87-a127-7c9f179d7272"
+_bugout_journal_id = "801e9b3c-6b03-40a7-870f-5b25d326da66"
+_telemetry_docs_url = docs_url("anonymous-telemetry")
+_telemetry_docs_referral = f"See {_telemetry_docs_url} for details"
+
+
+class AnonymousTelemetry(Subsystem):
+    options_scope = "anonymous-telemetry"
+    help = "Options related to sending anonymous stats to the Pants project, to aid development."
+
+    @classmethod
+    def register_options(cls, register):
+        register(
+            "--enabled",
+            advanced=True,
+            type=bool,
+            default=False,
+            help=(
+                f"Whether to send anonymous telemetry to the Pants project.\nTelemetry is sent "
+                f"asynchronously, with silent failure, and does not impact build times or "
+                f"outcomes.\n{_telemetry_docs_referral}."
+            ),
+        )
+        register(
+            "--repo-id",
+            advanced=True,
+            type=str,
+            default=None,
+            help=(
+                f"An anonymized ID representing this repo.\nFor private repos, you likely want the "
+                f"ID to not be derived from, or algorithmically convertible to, anything "
+                f"identifying the repo.\nFor public repos the ID may be visible in that repo's "
+                f"config file, so anonymity of the repo is not guaranteed (although user anonymity "
+                f"is always guaranteed).\n{_telemetry_docs_referral}."
+            ),
+        )
+
+    @property
+    def enabled(self) -> bool:
+        return cast(bool, self.options.enabled)
+
+    @property
+    def repo_id(self) -> str | None:
+        return cast("str | None", self.options.repo_id)
+
+
+class AnonymousTelemetryCallback(WorkunitsCallback):
+    def __init__(self, anonymous_telemetry: AnonymousTelemetry) -> None:
+        super().__init__()
+        self._anonymous_telemetry = anonymous_telemetry
+
+    @property
+    def can_finish_async(self) -> bool:
+        # Because we don't log anything, it's safe to finish in the background.
+        return True
+
+    @staticmethod
+    def validate_repo_id(repo_id: str) -> bool:
+        is_valid = re.match(r"^[a-zA-Z0-9-_]{30,60}$", repo_id) is not None
+        if not is_valid:
+            logger.error(
+                "The repo_id must be between 30 and 60 characters long, and consist of only "
+                "alphanumeric characters, dashes and underscores."
+            )
+        return is_valid
+
+    def __call__(
+        self,
+        *,
+        started_workunits: tuple[Workunit, ...],
+        completed_workunits: tuple[Workunit, ...],
+        finished: bool,
+        context: StreamingWorkunitContext,
+    ) -> None:
+        if not finished:
+            return
+
+        if self._anonymous_telemetry.options.is_default("enabled"):
+            logger.warning(
+                f"Please either set `enabled = true` in the [anonymous-telemetry] section of "
+                f"pants.toml to enable sending anonymous stats to the Pants project to aid "
+                f"development, or set `enabled = false` to disable it. No telemetry sent "
+                f"for this run. An explicit setting will get rid of this message. "
+                f"{_telemetry_docs_referral}."
+            )
+
+        if self._anonymous_telemetry.enabled:
+            repo_id = self._anonymous_telemetry.repo_id
+            if repo_id is None:
+                logger.error(
+                    f'Please set `repo_id = "<uuid>"` in the [anonymous-telemetry] section '
+                    f"of pants.toml, where `<uuid>` is some fixed random identifier, such as "
+                    f"one generated by uuidgen. No telemetry sent for this run. "
+                    f"{_telemetry_docs_referral}."
+                )
+            elif self.validate_repo_id(repo_id):
+                # Assemble and send the telemetry.
+                # Note that this method is called with finished=True only after the
+                # StreamingWorkunitHandler context ends, i.e., after end_run() has been called,
+                # so the RunTracker will have had a chance to finalize its state.
+                telemetry_data = context.run_tracker.get_anonymous_telemetry_data(repo_id)
+                # TODO: Add information about any errors that occurred.
+
+                reporter = Reporter(
+                    name="pantsbuild/pants",
+                    # We've already established consent at this point.
+                    consent=HumbugConsent(True),
+                    session_id=telemetry_data.get("run_id", str(uuid.uuid4())),
+                    bugout_token=_bugout_access_token,
+                    bugout_journal_id=_bugout_journal_id,
+                    timeout_seconds=5,
+                    # We don't want to spawn a thread in the engine, and we're
+                    # already running in a background thread in pantsd.
+                    mode=Modes.SYNCHRONOUS,
+                )
+
+                # This is copied from humbug code, to ensure that future changes to humbug
+                # don't add tags that inadvertently violate our anonymity promise.
+                system_tags = [
+                    "humbug",
+                    "source:{}".format(reporter.name),
+                    "os:{}".format(reporter.system_information.os),
+                    "arch:{}".format(reporter.system_information.machine),
+                    "python:{}".format(reporter.system_information.python_version_major),
+                    "python:{}.{}".format(
+                        reporter.system_information.python_version_major,
+                        reporter.system_information.python_version_minor,
+                    ),
+                    "python:{}".format(reporter.system_information.python_version),
+                    "session:{}".format(reporter.session_id),
+                ]
+                tags = (
+                    system_tags
+                    + [
+                        f"pants_version:{telemetry_data.get('pants_version')}",
+                    ]
+                    + [f"goal:{goal}" for goal in telemetry_data.get("goals", [])]
+                )
+
+                report = Report(
+                    title=f"pants run {reporter.session_id}",
+                    tags=tags,
+                    content=json.dumps(telemetry_data, sort_keys=True),
+                )
+                reporter.publish(report)
+
+
+class AnonymousTelemetryCallbackFactoryRequest:
+    """A unique request type that is installed to trigger construction of the WorkunitsCallback."""
+
+
+@rule
+def construct_callback(
+    _: AnonymousTelemetryCallbackFactoryRequest, anonymous_telemetry: AnonymousTelemetry
+) -> WorkunitsCallbackFactory:
+    return WorkunitsCallbackFactory(lambda: AnonymousTelemetryCallback(anonymous_telemetry))
+
+
+def rules():
+    return [
+        UnionRule(WorkunitsCallbackFactoryRequest, AnonymousTelemetryCallbackFactoryRequest),
+        *collect_rules(),
+    ]

src/python/pants/goal/anonymous_telemetry_integration_test.py

@@ -0,0 +1,60 @@
+# Copyright 2021 Pants project contributors (see CONTRIBUTORS.md).
+# Licensed under the Apache License, Version 2.0 (see LICENSE).
+
+from __future__ import annotations
+
+from pants.testutil.pants_integration_test import run_pants
+
+_no_explicit_setting_msg = "An explicit setting will get rid of this message"
+_no_repo_id_msg = 'set `repo_id = "<uuid>"` in the [anonymous-telemetry] section of pants.toml'
+_bad_repo_id_msg = "The repo_id must be between 30 and 60 characters long"
+
+
+def test_warn_if_no_explicit_setting() -> None:
+    result = run_pants(["roots"], config={}, use_pantsd=False)
+    result.assert_success()
+    assert _no_explicit_setting_msg in result.stderr
+    assert _no_repo_id_msg not in result.stderr
+    assert _bad_repo_id_msg not in result.stderr
+
+
+def test_warn_if_repo_id_unset() -> None:
+    result = run_pants(
+        ["roots"], config={"anonymous-telemetry": {"enabled": True}}, use_pantsd=False
+    )
+    result.assert_success()
+    assert _no_explicit_setting_msg not in result.stderr
+    assert _no_repo_id_msg in result.stderr
+    assert _bad_repo_id_msg not in result.stderr
+
+
+def test_warn_if_repo_id_invalid() -> None:
+    result = run_pants(
+        ["roots"],
+        config={"anonymous-telemetry": {"enabled": True, "repo_id": "tooshort"}},
+        use_pantsd=False,
+    )
+    result.assert_success()
+    assert _no_explicit_setting_msg not in result.stderr
+    assert _no_repo_id_msg not in result.stderr
+    assert _bad_repo_id_msg in result.stderr
+
+
+def test_no_warn_if_explicitly_on() -> None:
+    result = run_pants(
+        ["roots"],
+        config={"anonymous-telemetry": {"enabled": True, "repo_id": 36 * "a"}},
+        use_pantsd=False,
+    )
+    result.assert_success()
+    assert _no_explicit_setting_msg not in result.stderr
+    assert _no_repo_id_msg not in result.stderr
+    assert _bad_repo_id_msg not in result.stderr
+
+
+def test_no_warn_if_explicitly_off() -> None:
+    result = run_pants(["roots"], config={"anonymous-telemetry": {"enabled": False}})
+    result.assert_success()
+    assert _no_explicit_setting_msg not in result.stderr
+    assert _no_repo_id_msg not in result.stderr
+    assert _bad_repo_id_msg not in result.stderr

src/python/pants/goal/anonymous_telemetry_test.py

@@ -0,0 +1,35 @@
+# Copyright 2021 Pants project contributors (see CONTRIBUTORS.md).
+# Licensed under the Apache License, Version 2.0 (see LICENSE).
+
+import pytest
+
+from pants.goal.anonymous_telemetry import AnonymousTelemetryCallback
+
+
+@pytest.mark.parametrize(
+    "repo_id",
+    [
+        "a" * 30,
+        "2" * 31,
+        "C" * 60,
+        "c1db8737-06b4-4aa8-b18f-8cde023eb524",
+        "D2E39BA4_BA82_4A85_99DC_9E99E4528D3F",
+    ],
+)
+def test_valid_repo_ids(repo_id) -> None:
+    assert AnonymousTelemetryCallback.validate_repo_id(repo_id)
+
+
+@pytest.mark.parametrize(
+    "repo_id",
+    [
+        "",
+        "x",
+        "a" * 29,
+        "2" * 61,
+        "@c1db8737-06b4-4aa8-b18f-8cde023eb524",
+        "D2E39BA4-BA82-4A85-99DC-9Eá9E4528D3F",
+    ],
+)
+def test_invalid_repo_ids(repo_id) -> None:
+    assert not AnonymousTelemetryCallback.validate_repo_id(repo_id)