Skip to content

:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.

feat(data-warehouse): add 4 missing tables to the Buildkite source (#119962)

Tom Owers119bb1ddf044
diff --git a/products/warehouse_sources/backend/temporal/data_imports/sources/COVERAGE_GAPS_APPENDIX.md b/products/warehouse_sources/backend/temporal/data_imports/sources/COVERAGE_GAPS_APPENDIX.md
index cd80937daaad..4e26e96a2e95 100644
--- a/products/warehouse_sources/backend/temporal/data_imports/sources/COVERAGE_GAPS_APPENDIX.md
+++ b/products/warehouse_sources/backend/temporal/data_imports/sources/COVERAGE_GAPS_APPENDIX.md
@@ -2021,11 +2021,11 @@ Diffed against: <https://buildkite.com/docs/llms.txt>
- [x] `organizations/{org}/teams/{team}/pipelines` — membership table joining teams to pipelines, needed to attribute build cost per team (high)
- [x] `organizations/{org}/members` — lookup resolving the user IDs that appear as build creators and job agents (high)
- [x] `organizations/{org}/clusters/{cluster}/queues` — lookup resolving the queue an agent or job ran on — required for agent capacity and cost analysis (medium)
-- [ ] `organizations/{org}/clusters` — lookup resolving cluster IDs carried on agents and queues (medium)
-- [ ] `organizations/{org}/pipelines/{pipeline}/builds/{number}/artifacts` — artifact inventory and sizes per build, useful for storage and output tracking (medium)
-- [ ] `organizations/{org}/teams/{team}/members` — membership table joining users to teams (medium)
+- [x] `organizations/{org}/clusters` — lookup resolving cluster IDs carried on agents and queues (medium)
+- [x] `organizations/{org}/pipelines/{pipeline}/builds/{number}/artifacts` — artifact inventory and sizes per build, useful for storage and output tracking (medium)
+- [x] `organizations/{org}/teams/{team}/members` — membership table joining users to teams (medium)
- [x] `organizations/{org}/pipelines/{pipeline}/schedules` — explains which builds are scheduled vs triggered — a common breakdown dimension on build volume (medium)
-- [ ] `organizations/{org}/pipelines/{pipeline}/builds/{number}/annotations` — build annotations carry test summaries and custom CI reporting output (low)
+- [x] `organizations/{org}/pipelines/{pipeline}/builds/{number}/annotations` — build annotations carry test summaries and custom CI reporting output (low)
Note: Buildkite publishes a complete machine-readable docs index at /docs/llms.txt with one line per REST resource; every endpoint below is named there.
diff --git a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/buildkite.py b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/buildkite.py
index 74853c940ebf..a8e437ceaf70 100644
--- a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/buildkite.py
+++ b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/buildkite.py
@@ -22,6 +22,7 @@ from products.warehouse_sources.backend.temporal.data_imports.sources.common.res
from products.warehouse_sources.backend.temporal.data_imports.sources.common.rest_source.resource import Resource
from products.warehouse_sources.backend.temporal.data_imports.sources.common.rest_source.typing import (
ClientConfig,
+ Endpoint,
EndpointResource,
)
from products.warehouse_sources.backend.temporal.data_imports.sources.common.resumable import ResumableSourceManager
@@ -35,9 +36,12 @@ PAGE_SIZE = 100
# Test Engine suites are the fan-out parent for both Test Engine child tables.
SUITE_CHILD_ENDPOINTS = ("test_suite_runs", "test_suite_tests")
-# Clusters are only a fan-out parent, so the cluster list has no entry in the endpoint catalog and
-# is not offered as a table of its own.
-CLUSTERS_PATH = "/v2/organizations/{organization}/clusters"
+# Builds are the fan-out parent for every per-build child table.
+BUILD_CHILD_ENDPOINTS = ("jobs", "artifacts", "annotations")
+
+# Without `scope=all` the annotations listing returns only build-scoped annotations and drops the
+# job-scoped ones.
+BUILD_CHILD_EXTRA_PARAMS: dict[str, dict[str, Any]] = {"annotations": {"scope": "all"}}
@frozen
@@ -56,7 +60,7 @@ class BuildkiteFanoutConfig:
FANOUT_ENDPOINTS: dict[str, BuildkiteFanoutConfig] = {
"cluster_queues": BuildkiteFanoutConfig(
parent_name="clusters",
- parent_path=CLUSTERS_PATH,
+ parent_path=BUILDKITE_ENDPOINTS["clusters"].path,
resolve_param="cluster_id",
resolve_field="id",
parent_field_renames={"id": "cluster_id"},
@@ -75,6 +79,13 @@ FANOUT_ENDPOINTS: dict[str, BuildkiteFanoutConfig] = {
resolve_field="id",
parent_field_renames={"id": "team_id", "slug": "team_slug"},
),
+ "team_members": BuildkiteFanoutConfig(
+ parent_name="teams",
+ parent_path=BUILDKITE_ENDPOINTS["teams"].path,
+ resolve_param="team_id",
+ resolve_field="id",
+ parent_field_renames={"id": "team_id", "slug": "team_slug"},
+ ),
}
@@ -129,10 +140,10 @@ def _build_initial_params(
def _build_created_from(should_use_incremental_field: bool, db_incremental_field_last_value: Any) -> str | None:
- """Window the parent build walk that drives the jobs fan-out.
+ """Window the parent build walk that drives a per-build fan-out.
- Jobs have no time filter of their own, so the cursor a job row carries is its build's
- creation time and the window is applied to the builds listing that produced it.
+ Build children have no time filter of their own, so the cursor a child row carries is its
+ build's creation time and the window is applied to the builds listing that produced it.
"""
if not should_use_incremental_field or not db_incremental_field_last_value:
return None
@@ -230,7 +241,7 @@ def _fanout_child_resource(endpoint: str, organization: str) -> EndpointResource
def _flatten_pipeline_slug(row: dict[str, Any]) -> dict[str, Any]:
- # The jobs path needs the pipeline slug, which a build nests under `pipeline`; flatten it so
+ # A build child path needs the pipeline slug, which a build nests under `pipeline`; flatten it so
# both the path binding and include_from_parent can read it off the parent row.
pipeline = row.get("pipeline")
row["pipeline_slug"] = pipeline.get("slug") if isinstance(pipeline, dict) else None
@@ -246,25 +257,28 @@ def _builds_parent_resource(organization: str, created_from: str | None) -> Endp
return resource
-def _jobs_child_resource(organization: str) -> EndpointResource:
+def _build_child_resource(endpoint: str, organization: str) -> EndpointResource:
+ endpoint_config: Endpoint = {
+ "path": _bind_organization(BUILDKITE_ENDPOINTS[endpoint].path, organization),
+ "params": {
+ # Both path params bind fields of the same parent build row.
+ "pipeline_slug": {"type": "resolve", "resource": "builds", "field": "pipeline_slug"},
+ "build_number": {"type": "resolve", "resource": "builds", "field": "number"},
+ "per_page": PAGE_SIZE,
+ **BUILD_CHILD_EXTRA_PARAMS.get(endpoint, {}),
+ },
+ "data_selector_required": True,
+ "response_actions": [{"status_code": 404, "action": "ignore"}],
+ }
+ if endpoint == "jobs":
+ # Unlike the other list endpoints, jobs answers with an envelope and paginates on a
+ # cursor URL in the body rather than a Link header.
+ endpoint_config["data_selector"] = "items"
+ endpoint_config["paginator"] = JSONResponsePaginator(next_url_path="links.next")
return {
- "name": "jobs",
+ "name": endpoint,
"include_from_parent": ["id", "number", "pipeline_slug", "created_at"],
- "endpoint": {
- "path": _bind_organization(BUILDKITE_ENDPOINTS["jobs"].path, organization),
- "params": {
- # Both path params bind fields of the same parent build row.
- "pipeline_slug": {"type": "resolve", "resource": "builds", "field": "pipeline_slug"},
- "build_number": {"type": "resolve", "resource": "builds", "field": "number"},
- "per_page": PAGE_SIZE,
- },
- # Unlike the other list endpoints, jobs answers with an envelope and paginates on a
- # cursor URL in the body rather than a Link header.
- "data_selector": "items",
- "data_selector_required": True,
- "paginator": JSONResponsePaginator(next_url_path="links.next"),
- "response_actions": [{"status_code": 404, "action": "ignore"}],
- },
+ "endpoint": endpoint_config,
"data_map": rename_parent_fields(
"builds",
{
@@ -388,12 +402,15 @@ def buildkite_source(
_fanout_parent_resource(FANOUT_ENDPOINTS[endpoint], organization),
_fanout_child_resource(endpoint, organization),
]
- elif endpoint == "jobs":
+ elif endpoint in BUILD_CHILD_ENDPOINTS:
created_from = _build_created_from(should_use_incremental_field, db_incremental_field_last_value)
- resources = [_builds_parent_resource(organization, created_from), _jobs_child_resource(organization)]
+ resources = [
+ _builds_parent_resource(organization, created_from),
+ _build_child_resource(endpoint, organization),
+ ]
# Resume is deliberately off here: the framework rewrites the full set of completed child
# paths on every parent, which an organization's build list makes quadratic. A retry
- # re-walks the window and merge dedupes on the job id.
+ # re-walks the window and merge dedupes on the primary key.
manager = None
else:
params = _build_initial_params(
diff --git a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/canonical_descriptions.py b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/canonical_descriptions.py
index 572c14aa03b6..b4e1f5b55c64 100644
--- a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/canonical_descriptions.py
+++ b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/canonical_descriptions.py
@@ -97,6 +97,46 @@ CANONICAL_DESCRIPTIONS: CanonicalDescriptions = {
"source": "What triggered the build (e.g. ui, api, webhook, schedule).",
},
},
+ "artifacts": {
+ "description": "A file an agent uploaded during one of a build's jobs, with its path, type and size.",
+ "docs_url": "https://buildkite.com/docs/apis/rest-api/artifacts",
+ "columns": {
+ "id": "Unique identifier for the artifact.",
+ "job_id": "Identifier of the job that uploaded the artifact.",
+ "url": "API URL for the artifact.",
+ "download_url": "API URL that downloads the artifact contents.",
+ "state": "State of the artifact: new, error, finished, deleted, or expired.",
+ "path": "Path of the artifact.",
+ "dirname": "Path of the artifact without the filename.",
+ "filename": "Filename of the artifact.",
+ "mime_type": "MIME type of the artifact.",
+ "file_size": "File size of the artifact in bytes.",
+ "sha1sum": "SHA-1 hash of the artifact contents, as calculated by the agent.",
+ "build_id": "Identifier of the build the artifact belongs to.",
+ "build_number": "Number of the build the artifact belongs to.",
+ "build_created_at": "Time the parent build was created (ISO 8601).",
+ "pipeline_slug": "Slug of the pipeline the artifact's build ran for.",
+ },
+ },
+ "annotations": {
+ "description": "A Markdown snippet a job added to a build page, often a test summary or custom CI report.",
+ "docs_url": "https://buildkite.com/docs/apis/rest-api/annotations",
+ "columns": {
+ "id": "Unique identifier for the annotation.",
+ "context": "Context given when the build was annotated. Only one annotation per build has a given context.",
+ "style": "Style of the annotation: success, info, warning, or error.",
+ "scope": "Scope of the annotation, either build or job.",
+ "job_id": "Identifier of the job the annotation belongs to, or null for a build-scoped annotation.",
+ "priority": "Display priority of the annotation, from 1 to 10. Higher values show first.",
+ "body_html": "Rendered HTML of the annotation body.",
+ "created_at": "Time the annotation was first created (ISO 8601).",
+ "updated_at": "Time the annotation was last added to or replaced (ISO 8601).",
+ "build_id": "Identifier of the build the annotation belongs to.",
+ "build_number": "Number of the build the annotation belongs to.",
+ "build_created_at": "Time the parent build was created (ISO 8601).",
+ "pipeline_slug": "Slug of the pipeline the annotation's build ran for.",
+ },
+ },
"agents": {
"description": "A connected Buildkite agent that runs jobs for the organization.",
"docs_url": "https://buildkite.com/docs/apis/rest-api/agents",
@@ -154,6 +194,26 @@ CANONICAL_DESCRIPTIONS: CanonicalDescriptions = {
"pipeline_slug": "Slug of the pipeline the job's build ran for.",
},
},
+ "clusters": {
+ "description": "An isolated group of agents, queues and pipelines within the organization.",
+ "docs_url": "https://buildkite.com/docs/apis/rest-api/clusters",
+ "columns": {
+ "id": "Unique identifier for the cluster.",
+ "graphql_id": "GraphQL identifier for the cluster.",
+ "default_queue_id": "Identifier of the queue that agents join when they connect without a queue.",
+ "name": "Display name of the cluster.",
+ "description": "Human-readable description of the cluster.",
+ "emoji": "Emoji for the cluster, in Buildkite emoji syntax.",
+ "color": "Color hex code for the cluster.",
+ "maintainers": "The users and teams that maintain the cluster.",
+ "url": "API URL for the cluster.",
+ "web_url": "URL of the cluster in the Buildkite web UI.",
+ "queues_url": "API URL for the cluster's queues.",
+ "default_queue_url": "API URL for the cluster's default queue.",
+ "created_at": "Time the cluster was created (ISO 8601).",
+ "created_by": "The user who created the cluster.",
+ },
+ },
"cluster_queues": {
"description": "A queue within a cluster, resolving the queue an agent or job ran on for capacity and cost analysis.",
"docs_url": "https://buildkite.com/docs/apis/rest-api/clusters/queues",
@@ -201,6 +261,18 @@ CANONICAL_DESCRIPTIONS: CanonicalDescriptions = {
"team_slug": "Slug of the team the pipeline is linked to.",
},
},
+ "team_members": {
+ "description": "A link between a team and a user, with the role the user holds in the team.",
+ "docs_url": "https://buildkite.com/docs/apis/rest-api/teams/members",
+ "columns": {
+ "user_id": "UUID of the user.",
+ "user_name": "Name of the user.",
+ "role": "Role the user holds in the team, either member or maintainer.",
+ "created_at": "Time the user was added to the team (ISO 8601).",
+ "team_id": "Identifier of the team the user belongs to.",
+ "team_slug": "Slug of the team the user belongs to.",
+ },
+ },
"test_suites": {
"description": "A Test Engine suite collecting the test results uploaded for a project.",
"docs_url": "https://buildkite.com/docs/apis/rest-api/test-engine/suites",
diff --git a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/settings.py b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/settings.py
index f44beb34d32d..620a24f1b1ad 100644
--- a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/settings.py
+++ b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/settings.py
@@ -34,8 +34,8 @@ class BuildkiteEndpointConfig:
# Buildkite v2 REST API endpoints. Most resources are nested under an organization slug, and the
# Test Engine resources live under the /v2/analytics prefix. Only the builds endpoint exposes a
-# server-side timestamp filter (created_from); jobs inherit it through their parent build fan-out,
-# and the rest are full refresh.
+# server-side timestamp filter (created_from); jobs, artifacts and annotations inherit it through
+# their parent build fan-out, and the rest are full refresh.
BUILDKITE_ENDPOINTS: dict[str, BuildkiteEndpointConfig] = {
"organizations": BuildkiteEndpointConfig(
name="organizations",
@@ -115,12 +115,67 @@ BUILDKITE_ENDPOINTS: dict[str, BuildkiteEndpointConfig] = {
"Disabled by default because of the API cost"
),
),
+ "artifacts": BuildkiteEndpointConfig(
+ name="artifacts",
+ path="/v2/organizations/{organization}/pipelines/{pipeline_slug}/builds/{build_number}/artifacts",
+ # Buildkite only addresses an artifact through its build and job, so the build is keyed on.
+ primary_keys=["build_id", "id"],
+ # Artifacts have no time filter or timestamp of their own, so they share the jobs window.
+ partition_key="build_created_at",
+ incremental_fields=[
+ {
+ "label": "build_created_at",
+ "type": IncrementalFieldType.DateTime,
+ "field": "build_created_at",
+ "field_type": IncrementalFieldType.DateTime,
+ },
+ ],
+ sort_mode="desc",
+ probe_path="/v2/organizations/{organization}/builds",
+ # An artifact's state moves from new to finished, and later to expired or deleted.
+ merge_only=True,
+ should_sync_default=False,
+ description=(
+ "Fetches the artifacts of every build, at least one request per build. "
+ "Disabled by default because of the API cost"
+ ),
+ ),
+ "annotations": BuildkiteEndpointConfig(
+ name="annotations",
+ path="/v2/organizations/{organization}/pipelines/{pipeline_slug}/builds/{build_number}/annotations",
+ primary_keys=["build_id", "id"],
+ # `created_at` is the annotation's own, but the server-side window applies to the parent build.
+ partition_key="build_created_at",
+ incremental_fields=[
+ {
+ "label": "build_created_at",
+ "type": IncrementalFieldType.DateTime,
+ "field": "build_created_at",
+ "field_type": IncrementalFieldType.DateTime,
+ },
+ ],
+ sort_mode="desc",
+ probe_path="/v2/organizations/{organization}/builds",
+ # A step can append to or replace an annotation after it is created.
+ merge_only=True,
+ should_sync_default=False,
+ description=(
+ "Fetches the annotations of every build, at least one request per build. "
+ "Disabled by default because of the API cost"
+ ),
+ ),
"agents": BuildkiteEndpointConfig(
name="agents",
path="/v2/organizations/{organization}/agents",
partition_key="created_at",
incremental_fields=[],
),
+ "clusters": BuildkiteEndpointConfig(
+ name="clusters",
+ path="/v2/organizations/{organization}/clusters",
+ partition_key="created_at",
+ incremental_fields=[],
+ ),
"cluster_queues": BuildkiteEndpointConfig(
name="cluster_queues",
path="/v2/organizations/{organization}/clusters/{cluster_id}/queues",
@@ -148,6 +203,16 @@ BUILDKITE_ENDPOINTS: dict[str, BuildkiteEndpointConfig] = {
incremental_fields=[],
probe_path="/v2/organizations/{organization}/teams",
),
+ "team_members": BuildkiteEndpointConfig(
+ name="team_members",
+ path="/v2/organizations/{organization}/teams/{team_id}/members",
+ # The rows are team-to-user links with no id of their own, so the pair they join is the key.
+ primary_keys=["team_id", "user_id"],
+ # `created_at` is when the user joined the team, which is set once.
+ partition_key="created_at",
+ incremental_fields=[],
+ probe_path="/v2/organizations/{organization}/teams",
+ ),
"test_suites": BuildkiteEndpointConfig(
name="test_suites",
path="/v2/analytics/organizations/{organization}/suites",
diff --git a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/source.py b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/source.py
index e21e7d668c4d..14ba165d36ad 100644
--- a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/source.py
+++ b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/source.py
@@ -87,6 +87,7 @@ Make sure to grant the following read scopes:
- `read_organizations`
- `read_pipelines`
- `read_builds`
+- `read_artifacts`
- `read_agents`
- `read_clusters`
- `read_teams`
diff --git a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/tests/test_buildkite.py b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/tests/test_buildkite.py
index df74747496e0..02dd1abffd20 100644
--- a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/tests/test_buildkite.py
+++ b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/tests/test_buildkite.py
@@ -199,6 +199,9 @@ class TestValidateCredentials:
("team_pipelines", f"{ORG_URL}/teams?per_page=1"),
("pipeline_schedules", f"{ORG_URL}/pipelines?per_page=1"),
("cluster_queues", f"{ORG_URL}/clusters?per_page=1"),
+ ("team_members", f"{ORG_URL}/teams?per_page=1"),
+ ("artifacts", f"{ORG_URL}/builds?per_page=1"),
+ ("annotations", f"{ORG_URL}/builds?per_page=1"),
]
)
def test_fanout_schema_probe_targets_the_parent_listing(self, endpoint: str, expected_url: str) -> None:
@@ -268,6 +271,14 @@ class TestSimpleFanout:
[{"id": "s1"}, {"id": "s2"}],
[{"pipeline_slug": "web"}, {"pipeline_slug": "api"}],
),
+ (
+ "team_members",
+ [{"id": "t1", "slug": "backend"}, {"id": "t2", "slug": "frontend"}],
+ f"{ORG_URL}/teams",
+ [f"{ORG_URL}/teams/t1/members", f"{ORG_URL}/teams/t2/members"],
+ [{"user_id": "u1", "role": "member"}, {"user_id": "u1", "role": "maintainer"}],
+ [{"team_id": "t1", "team_slug": "backend"}, {"team_id": "t2", "team_slug": "frontend"}],
+ ),
(
"cluster_queues",
[{"id": "c1"}, {"id": "c2"}],
@@ -330,3 +341,46 @@ class TestJobsFanout:
assert result.requests[0].param("created_from") == "2026-03-04T02:58:14+00:00"
assert result.requests[0].param("per_page") == "100"
assert result.requests[1].param("created_from") is None
+
+
+class TestBuildChildFanout:
+ @parameterized.expand(
+ [
+ ("artifacts", {}),
+ # Without scope=all the listing drops every job-scoped annotation.
+ ("annotations", {"scope": ("all",)}),
+ ]
+ )
+ def test_walks_every_page_of_each_build_inside_the_incremental_window(
+ self, endpoint: str, extra_query: dict[str, tuple[str, ...]]
+ ) -> None:
+ child_base = f"{ORG_URL}/pipelines/web/builds/42/{endpoint}"
+ page2 = f"{child_base}?page=2&per_page=100"
+ result = _driver().run(
+ endpoint,
+ [
+ _response(
+ [{"id": "b1", "number": 42, "created_at": "2026-03-04T02:58:14Z", "pipeline": {"slug": "web"}}]
+ ),
+ # Unlike jobs, these children are top-level arrays paginated on a Link header.
+ _response([{"id": "x1"}], link=page2),
+ _response([{"id": "x2"}]),
+ ],
+ should_use_incremental_field=True,
+ db_incremental_field_last_value=datetime(2026, 3, 4, 2, 58, 14, tzinfo=UTC),
+ incremental_field="build_created_at",
+ )
+
+ assert result.raised is None
+ assert result.requests[0].param("created_from") == "2026-03-04T02:58:14+00:00"
+ assert result.urls[1].split("?")[0] == child_base
+ assert result.queries[1] == {"per_page": ("100",), **extra_query}
+ assert result.urls[2] == page2
+ # The build id is part of the primary key, and build_created_at is the incremental cursor.
+ assert [
+ (row["id"], row["build_id"], row["build_number"], row["pipeline_slug"], row["build_created_at"])
+ for row in result.rows
+ ] == [
+ ("x1", "b1", 42, "web", "2026-03-04T02:58:14Z"),
+ ("x2", "b1", 42, "web", "2026-03-04T02:58:14Z"),
+ ]
diff --git a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/tests/test_buildkite_source.py b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/tests/test_buildkite_source.py
index 3e51a7da826a..6b3f4687c977 100644
--- a/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/tests/test_buildkite_source.py
+++ b/products/warehouse_sources/backend/temporal/data_imports/sources/buildkite/tests/test_buildkite_source.py
@@ -31,6 +31,8 @@ class TestBuildkiteSource:
# fan-out that drives them. Nothing else has one.
("builds", True, True),
("jobs", True, False),
+ ("artifacts", True, False),
+ ("annotations", True, False),
("organizations", False, False),
("organization_members", False, False),
("pipelines", False, False),
@@ -47,8 +49,8 @@ class TestBuildkiteSource:
def test_incremental_support_per_endpoint(self, endpoint: str, incremental: bool, append: bool) -> None:
schemas = {s.name: s for s in self.source.get_schemas(_config(), team_id=self.team_id)}
assert schemas[endpoint].supports_incremental is incremental
- # A job restates after it is created (state, finished_at) and each incremental run re-walks
- # the trailing build window, so append would materialize the re-pulled rows as duplicates.
+ # A build child restates after it is created (a job's state, an artifact's state, an
+ # annotation's body), so append would materialize the re-pulled rows as duplicates.
assert schemas[endpoint].supports_append is append
@parameterized.expand(