CLOUDP-305848 & CLOUDP-338152 - release on master merges axuillary images

.evergreen-functions.yml

@@ -496,6 +496,7 @@ functions:
           IMAGE_NAME: ${image_name}
           BUILD_SCENARIO_OVERRIDE: ${build_scenario}
           FLAGS: ${flags}
+          AGENT_VERSION_OVERRIDE: ${agent_version}
 
   teardown_cloud_qa_all:
     - command: shell.exec

.evergreen.yml

@@ -287,19 +287,21 @@ tasks:
     commands:
       - func: lint_repo
 
-  # pct only triggers this variant once a new agent image is out
-  - name: release_agent
-    # this enables us to run this variant either manually (patch) which pct does or during an OM bump (github_pr)
-    allowed_requesters: [ "patch", "github_pr" ]
+  # Runs on every merge - detects release.json changes and releases appropriate images
+  - name: release_om_and_agents
+    allowed_requesters: [ "patch" , "commit"]
     commands:
       - func: clone
+      - func: python_venv
       - func: setup_building_host
       - func: quay_login
       - func: setup_docker_sbom
-      - func: pipeline
-        vars:
-          image_name: agent
-          build_scenario: release
+      - command: subprocess.exec
+        params:
+          working_dir: src/github.com/mongodb/mongodb-kubernetes
+          binary: scripts/dev/run_python.sh
+          args:
+            - scripts/release/release_om_and_agents.py
 
   - name: migrate_all_agents
     # this enables us to run this variant manually to build all the agents for the new agent registry
@@ -343,8 +345,9 @@ tasks:
       - func: quay_login
       - func: pipeline
         vars:
+          agent_version: all
           image_name: agent
-          flags: "--parallel --all-agents --skip-if-exists=false"
+          flags: "--parallel --skip-if-exists=false"
 
   - name: rebuild_currently_used_agents
     # this enables us to run this manually (patch) and rebuild current agent versions to verify
@@ -356,8 +359,9 @@ tasks:
       - func: quay_login
       - func: pipeline
         vars:
+          agent_version: current
           image_name: agent
-          flags: "--parallel --current-agents --skip-if-exists=false"
+          flags: "--parallel --skip-if-exists=false"
 
   - name: build_kubectl_mongodb_plugin
     commands:
@@ -463,7 +467,8 @@ tasks:
       - func: pipeline
         vars:
           image_name: agent
-          flags: "--parallel --all-agents"
+          agent_version: all
+          flags: "--parallel"
 
   - name: build_init_database_image_ubi
     commands:
@@ -749,7 +754,7 @@ task_groups:
       - e2e_sharded_cluster_scram_sha_256_switch_project
       - e2e_replica_set_scram_sha_1_switch_project
       - e2e_sharded_cluster_scram_sha_1_switch_project
-      # TODO CLOUDP-349093 - Disabled these tests as they don't use the password secret, and project migrations aren't fully supported yet.  
+      # TODO CLOUDP-349093 - Disabled these tests as they don't use the password secret, and project migrations aren't fully supported yet.
       # e2e_sharded_cluster_x509_switch_project
       # e2e_replica_set_x509_switch_project
       # e2e_replica_set_ldap_switch_project
@@ -1893,26 +1898,6 @@ buildvariants:
     tasks:
       - name: build_om_images
 
-    # It will be called by pct while bumping the agent cloud manager image
-  - name: release_agent
-    display_name: release_agent
-    tags: [ "manual_patch", "release_agent" ]
-    run_on:
-      - release-ubuntu2404-small # This is required for CISA attestation https://jira.mongodb.org/browse/DEVPROD-17780
-    depends_on:
-      - variant: init_test_run
-        name: build_agent_images_ubi # this ensures the agent gets released to ECR as well
-      - variant: e2e_multi_cluster_kind
-        name: '*'
-      - variant: e2e_static_multi_cluster_2_clusters
-        name: '*'
-      - variant: e2e_mdb_kind_ubi_cloudqa
-        name: '*'
-      - variant: e2e_static_mdb_kind_ubi_cloudqa
-        name: '*'
-    tasks:
-      - name: release_agent
-
     # Only called manually, It's used for testing the task release_agents in case the release.json
     # has not changed, and you still want to push the images to registry.
   - name: manual_release_all_agents
@@ -1946,51 +1931,6 @@ buildvariants:
       - name: backup_csv_images_limit_3
       - name: backup_csv_images_all
 
-  - name: publish_om60_images
-    display_name: publish_om60_images
-    tags: [ "manual_patch" ]
-    allowed_requesters: [ "patch", "github_pr" ]
-    run_on:
-      - release-ubuntu2404-small # This is required for CISA attestation https://jira.mongodb.org/browse/DEVPROD-17780
-    depends_on:
-      - variant: e2e_om60_kind_ubi
-        name: '*'
-      - variant: e2e_static_om60_kind_ubi
-        name: '*'
-    tasks:
-      - name: publish_ops_manager
-      - name: release_agent
-
-  - name: publish_om70_images
-    display_name: publish_om70_images
-    tags: [ "manual_patch" ]
-    allowed_requesters: [ "patch", "github_pr" ]
-    run_on:
-      - release-ubuntu2404-small # This is required for CISA attestation https://jira.mongodb.org/browse/DEVPROD-17780
-    depends_on:
-      - variant: e2e_om70_kind_ubi
-        name: '*'
-      - variant: e2e_static_om70_kind_ubi
-        name: '*'
-    tasks:
-      - name: publish_ops_manager
-      - name: release_agent
-
-  - name: publish_om80_images
-    display_name: publish_om80_images
-    tags: [ "manual_patch" ]
-    allowed_requesters: [ "patch", "github_pr" ]
-    run_on:
-      - release-ubuntu2404-small # This is required for CISA attestation https://jira.mongodb.org/browse/DEVPROD-17780
-    depends_on:
-      - variant: e2e_om80_kind_ubi
-        name: '*'
-      - variant: e2e_static_om80_kind_ubi
-        name: '*'
-    tasks:
-      - name: publish_ops_manager
-      - name: release_agent
-
   - name: migrate_all_agents
     display_name: migrate_all_agents
     tags: [ "manual_patch" ]
@@ -1999,3 +1939,13 @@ buildvariants:
       - ubuntu2404-large
     tasks:
       - name: migrate_all_agents
+
+  # Runs on every merge to master and releases images auxiliary to the operator release like OM and the Agent
+  - name: release_om_and_agents
+    display_name: release_om_and_agents
+    allowed_requesters: [ "patch" , "commit"]
+    run_on:
+      - release-ubuntu2404-small
+    patchable: false  # Only run on commit builds
+    tasks:
+      - name: release_om_and_agents

scripts/release/agent/detect_ops_manager_changes.py → scripts/release/agent/agents_to_rebuild.py

@@ -19,32 +19,6 @@
 logger = logging.getLogger(__name__)
 
 
-def get_content_from_git(commit: str, file_path: str) -> Optional[str]:
-    try:
-        result = subprocess.run(
-            ["git", "show", f"{commit}:{file_path}"], capture_output=True, text=True, check=True, timeout=30
-        )
-        return result.stdout
-    except (subprocess.CalledProcessError, subprocess.TimeoutExpired) as e:
-        logger.error(f"Failed to get {file_path} from git commit {commit}: {e}")
-        return None
-
-
-def load_release_json_from_master() -> Optional[Dict]:
-    base_revision = "origin/master"
-
-    content = get_content_from_git(base_revision, "release.json")
-    if not content:
-        logger.error(f"Could not retrieve release.json from {base_revision}")
-        return None
-
-    try:
-        return json.loads(content)
-    except json.JSONDecodeError as e:
-        logger.error(f"Invalid JSON in base release.json: {e}")
-        return None
-
-
 def load_current_release_json() -> Optional[Dict]:
     try:
         with open("release.json", "r") as f:
@@ -60,28 +34,6 @@ def extract_ops_manager_mapping(release_data: Dict) -> Dict:
     return release_data.get("supportedImages", {}).get("mongodb-agent", {}).get("opsManagerMapping", {})
 
 
-def get_changed_agents(current_mapping: Dict, base_mapping: Dict) -> List[Tuple[str, str]]:
-    """Returns list of (agent_version, tools_version) tuples for added/changed agents"""
-    added_agents = []
-
-    current_om_mapping = current_mapping.get("ops_manager", {})
-    master_om_mapping = base_mapping.get("ops_manager", {})
-
-    for om_version, agent_tools_version in current_om_mapping.items():
-        if om_version not in master_om_mapping or master_om_mapping[om_version] != agent_tools_version:
-            added_agents.append((agent_tools_version["agent_version"], agent_tools_version["tools_version"]))
-
-    current_cm = current_mapping.get("cloud_manager")
-    master_cm = base_mapping.get("cloud_manager")
-    current_cm_tools = current_mapping.get("cloud_manager_tools")
-    master_cm_tools = base_mapping.get("cloud_manager_tools")
-
-    if current_cm != master_cm or current_cm_tools != master_cm_tools:
-        added_agents.append((current_cm, current_cm_tools))
-
-    return list(set(added_agents))
-
-
 def get_tools_version_for_agent(agent_version: str) -> str:
     """Get tools version for a given agent version from release.json"""
     release_data = load_current_release_json()
@@ -206,26 +158,3 @@ def get_currently_used_agents() -> List[Tuple[str, str]]:
     except Exception as e:
         logger.error(f"Error getting currently used agents: {e}")
         return []
-
-
-def detect_ops_manager_changes() -> List[Tuple[str, str]]:
-    """Returns (has_changes, changed_agents_list)"""
-    logger.info("=== Detecting OM Mapping Changes (Local vs Base) ===")
-
-    current_release = load_current_release_json()
-    if not current_release:
-        logger.error("Could not load current local release.json")
-        return []
-
-    master_release = load_release_json_from_master()
-    if not master_release:
-        logger.warning("Could not load base release.json, assuming changes exist")
-        return []
-
-    current_mapping = extract_ops_manager_mapping(current_release)
-    base_mapping = extract_ops_manager_mapping(master_release)
-
-    if current_mapping != base_mapping:
-        return get_changed_agents(current_mapping, base_mapping)
-    else:
-        return []

scripts/release/atomic_pipeline.py

@@ -15,8 +15,7 @@
 from opentelemetry import trace
 
 from lib.base_logger import logger
-from scripts.release.agent.detect_ops_manager_changes import (
-    detect_ops_manager_changes,
+from scripts.release.agent.agents_to_rebuild import (
     get_all_agents_for_rebuild,
     get_currently_used_agents,
 )
@@ -330,22 +329,23 @@ def build_upgrade_hook_image(build_configuration: ImageBuildConfiguration):
 
 
 def build_agent(build_configuration: ImageBuildConfiguration):
-    """
-    Build the agent only for the latest operator for patches and operator releases.
+    """Build the agent image(s). Validation happens in pipeline.py."""
+    version = build_configuration.version
 
-    """
-    if build_configuration.all_agents:
+    if version == "all":
         agent_versions_to_build = get_all_agents_for_rebuild()
         logger.info("building all agents")
-    elif build_configuration.currently_used_agents:
+    elif version == "current":
         agent_versions_to_build = get_currently_used_agents()
-        logger.info("building current used agents")
+        logger.info("building currently used agents")
+    elif version and build_configuration.agent_tools_version:
+        agent_versions_to_build = [(version, build_configuration.agent_tools_version)]
+        logger.info(f"building agent {version} with tools {build_configuration.agent_tools_version}")
     else:
-        agent_versions_to_build = detect_ops_manager_changes()
-        logger.info("building agents for changed OM versions")
+        raise ValueError("No agent selection provided - this should be caught by pipeline.py validation")
 
     if not agent_versions_to_build:
-        logger.info("No changes detected, skipping agent build")
+        logger.warning("No agent versions found to build")
         return
 
     logger.info(f"Building Agent versions: {agent_versions_to_build}")

scripts/release/build/image_build_configuration.py

@@ -4,8 +4,7 @@
 from scripts.release.build.build_scenario import BuildScenario
 from scripts.release.build.image_build_process import ImageBuilder
 
-SUPPORTED_PLATFORMS = ["darwin/amd64", "darwin/arm64", "linux/amd64", "linux/arm64", "linux/s390x",
-                       "linux/ppc64le"]
+SUPPORTED_PLATFORMS = ["darwin/amd64", "darwin/arm64", "linux/amd64", "linux/arm64", "linux/s390x", "linux/ppc64le"]
 
 
 @dataclass
@@ -24,9 +23,8 @@ class ImageBuildConfiguration:
     # Agent specific
     parallel: bool = False
     parallel_factor: int = 0
-    all_agents: bool = False
-    currently_used_agents: bool = False
     architecture_suffix: bool = False
+    agent_tools_version: Optional[str] = None  # Explicit tools version for agent builds
 
     def is_release_scenario(self) -> bool:
         return self.scenario == BuildScenario.RELEASE

scripts/release/pipeline.py

@@ -62,6 +62,9 @@
 )
 from scripts.release.build.image_build_process import PodmanImageBuilder
 
+CURRENT_AGENTS = "current"
+ALL_AGENTS = "all"
+
 """
 The goal of main.py, image_build_configuration.py and build_context.py is to provide a single source of truth for the build
 configuration. All parameters that depend on the the build environment (local dev, evg, etc) should be resolved here and
@@ -139,9 +142,25 @@ def image_build_config_from_args(args) -> ImageBuildConfiguration:
     if type(builder) is PodmanImageBuilder and len(platforms) > 1:
         raise ValueError("Cannot use Podman builder with multi-platform builds")
 
-    # Validate version - only agent can have None version as the versions are managed by the agent
-    # which are externally retrieved from release.json
-    if version is None and image != "agent":
+    # Get agent_tools_version for agent builds (from --agent-tools-version arg)
+    agent_tools_version = getattr(args, "agent_tools_version", None)
+
+    # Validate version requirements
+    if image == "agent":
+        # Agent builds: version can be "all", "current", or explicit version (requires agent_tools_version)
+        if version is None:
+            raise ValueError(
+                "Agent build requires --version. Use one of:\n"
+                "  --version all                                      (for all agents in release.json)\n"
+                "  --version current                                  (for currently used agents)\n"
+                "  --version <ver> --agent-tools-version <tools_ver>  (for specific agent)"
+            )
+        is_special_version = version in (ALL_AGENTS, CURRENT_AGENTS)
+        if not is_special_version and agent_tools_version is None:
+            raise ValueError(
+                f"For agent builds with explicit version '{version}', --agent-tools-version must also be provided."
+            )
+    elif version is None:
         raise ValueError(f"Version cannot be empty for {image}.")
 
     return ImageBuildConfiguration(
@@ -157,9 +176,8 @@ def image_build_config_from_args(args) -> ImageBuildConfiguration:
         skip_if_exists=skip_if_exists,
         parallel=args.parallel,
         parallel_factor=args.parallel_factor,
-        all_agents=args.all_agents,
-        currently_used_agents=args.current_agents,
         architecture_suffix=architecture_suffix,
+        agent_tools_version=agent_tools_version,
     )
 
 
@@ -274,14 +292,11 @@ def main():
         help="Number of agent builds to run in parallel, defaults to number of cores",
     )
     parser.add_argument(
-        "--all-agents",
-        action="store_true",
-        help="Build all agent images.",
-    )
-    parser.add_argument(
-        "--current-agents",
-        action="store_true",
-        help="Build all currently used agent images.",
+        "--agent-tools-version",
+        metavar="",
+        action="store",
+        type=str,
+        help="Tools version to use when building agent image. Required when --version is an explicit version (not 'all' or 'current').",
     )
     parser.add_argument(
         "--architecture-suffix",