Archive Events in OpenSearch

This commit implements archiving of events in an OpenSearch backend.

It introduces a new subscriber group that reads from the events stream
and stores events in an OpenSearch instance.

Additionally, the worker pool now supports execution on specific task
queues, enabling different workers to process tasks based on their queue.

Signed-off-by: Jose Javier Merchante <[email protected]>

Author: jjmerchante

config/settings/testing.py

@@ -67,6 +67,7 @@ def __call__(self, _, strict):
         return self.conn
 
 
+RQ_QUEUES['default'] = _RQ_DATABASE  # noqa: F405
 RQ_QUEUES['testing'] = _RQ_DATABASE  # noqa: F405
 RQ['WORKER_CLASS'] = rq.worker.SimpleWorker
 

poetry.lock

@@ -54,6 +54,7 @@ perceval = {version = "^1.0.2", allow-prereleases = true}
 grimoirelab-chronicler = {git = "https://github.com/chaoss/grimoirelab-chronicler.git", allow-prereleases = true}
 django-cors-headers = "^4.6.0"
 djangorestframework = "^3.15.2"
+opensearch-py = "^2.8.0"
 
 [tool.poetry.group.dev.dependencies]
 fakeredis = "^2.0.0"

pyproject.toml

@@ -273,7 +273,8 @@
 # https://github.com/rq/django-rq
 #
 
-GRIMOIRELAB_Q_EVENTIZER_JOBS = os.environ.get('GRIMOIRELAB_Q_EVENTIZER_JOBS', 'default')
+GRIMOIRELAB_Q_EVENTIZER_JOBS = os.environ.get('GRIMOIRELAB_Q_EVENTIZER_JOBS', 'eventizer_jobs')
+GRIMOIRELAB_Q_ARCHIVIST_JOBS = os.environ.get('GRIMOIRELAB_Q_ARCHIVIST_JOBS', 'storage_jobs')
 
 _RQ_DATABASE = {
     'HOST': os.environ.get('GRIMOIRELAB_REDIS_HOST', '127.0.0.1'),
@@ -284,6 +285,7 @@
 
 RQ_QUEUES = {
     GRIMOIRELAB_Q_EVENTIZER_JOBS: _RQ_DATABASE,
+    GRIMOIRELAB_Q_ARCHIVIST_JOBS: _RQ_DATABASE,
 }
 
 GRIMOIRELAB_EVENTS_STREAM_NAME = os.environ.get('GRIMOIRELAB_EVENTS_STREAM_NAME',
@@ -309,3 +311,15 @@
 GRIMOIRELAB_JOB_TIMEOUT = int(os.environ.get('GRIMOIRELAB_JOB_TIMEOUT', -1))
 
 GRIMOIRELAB_GIT_STORAGE_PATH = os.environ.get('GRIMOIRELAB_GIT_PATH', '~/.perceval')
+
+#
+# Archivist configuration
+#
+GRIMOIRELAB_ARCHIVIST = {
+    'WORKERS': int(os.environ.get('GRIMOIRELAB_ARCHIVIST_WORKERS', 10)),
+    'STORAGE_TYPE': os.environ.get('GRIMOIRELAB_ARCHIVIST_STORAGE_TYPE', 'opensearch'),
+    'STORAGE_URL': os.environ.get('GRIMOIRELAB_ARCHIVIST_STORAGE_URL', 'https://admin:admin@localhost:9200'),
+    'STORAGE_INDEX': os.environ.get('GRIMOIRELAB_ARCHIVIST_STORAGE_INDEX', 'events'),
+    'STORAGE_VERIFY_CERT': os.environ.get('GRIMOIRELAB_ARCHIVIST_STORAGE_VERIFY_CERT', 'False').lower() in ('true', '1'),
+    'EVENTS_PER_JOB': int(os.environ.get('GRIMOIRELAB_ARCHIVIST_EVENTS_PER_JOB', 10000)),
+}

src/grimoirelab/core/config/settings.py

@@ -19,6 +19,7 @@
 from __future__ import annotations
 
 import os
+import sys
 import typing
 
 import click
@@ -43,9 +44,13 @@ def run(ctx: Context):
               is_flag=True,
               default=False,
               help="Run the service in developer mode.")
+@click.option("--clear-tasks",
+              is_flag=True,
+              default=False,
+              help="Clear background tasks.")
 @run.command()
 @click.pass_context
-def server(ctx: Context, devel: bool):
+def server(ctx: Context, devel: bool, clear_tasks: bool):
     """Start the GrimoireLab core server.
 
     GrimoireLab server allows to schedule tasks and fetch data from
@@ -56,6 +61,8 @@ def server(ctx: Context, devel: bool):
     should be run with a reverse proxy. If you activate the '--dev' flag,
     a HTTP server will be run instead.
     """
+    create_background_tasks(clear_tasks)
+
     env = os.environ
 
     env["UWSGI_ENV"] = f"DJANGO_SETTINGS_MODULE={ctx.obj['cfg']}"
@@ -91,24 +98,89 @@ def server(ctx: Context, devel: bool):
 
 
 @run.command()
+@click.argument('task-types', nargs=-1)
 @click.option('--workers',
               default=10,
               show_default=True,
               help="Number of workers to run in the pool.")
-def eventizers(workers: int):
-    """Start a pool of eventizer workers.
+def worker_pool(task_types: str, workers: int):
+    """Start a pool of workers that run specific tasks.
 
-    The workers on the pool will run tasks to fetch data from software
-    development repositories. Data will be processed in form of events,
-    and published in the events queue.
+    If multiple tasks share the same queue, they will run in the same
+    pool of workers. The tasks to run are defined as arguments to the
+    command.
 
     The number of workers running in the pool can be defined with the
     parameter '--workers'.
-
-    Workers get jobs from the GRIMOIRELAB_Q_EVENTIZER_JOBS queue defined
-    in the configuration file.
     """
+    from grimoirelab.core.scheduler.models import (get_registered_task_model,
+                                                   get_all_registered_task_names)
+
+    available_tasks = get_all_registered_task_names()
+
+    queues = []
+    for task in task_types:
+        try:
+            Task = get_registered_task_model(task)[0]
+        except KeyError:
+            click.echo(f"Task '{task}' is not a valid task. "
+                       f"Options: {available_tasks}", err=True)
+            sys.exit(1)
+        queues.append(Task().default_job_queue)
+
+    if not queues:
+        click.echo(f"You must define at least one valid task. "
+                   f"Options: {available_tasks}", err=True)
+        sys.exit(1)
+
     django.core.management.call_command(
-        'rqworker-pool', settings.GRIMOIRELAB_Q_EVENTIZER_JOBS,
+        'rqworker-pool', queues,
         num_workers=workers
     )
+
+
+def create_background_tasks(clear_tasks: bool):
+    """
+    Create background tasks before starting the server.
+    :param clear_tasks: clear tasks before creating new ones.
+    :return:
+    """
+    from grimoirelab.core.scheduler.scheduler import schedule_task
+    from grimoirelab.core.scheduler.tasks.models import StorageTask
+
+    workers = settings.GRIMOIRELAB_ARCHIVIST['WORKERS']
+    storage_url = settings.GRIMOIRELAB_ARCHIVIST['STORAGE_URL']
+    storage_db_name = settings.GRIMOIRELAB_ARCHIVIST['STORAGE_INDEX']
+    storage_type = settings.GRIMOIRELAB_ARCHIVIST['STORAGE_TYPE']
+    verify_certs = settings.GRIMOIRELAB_ARCHIVIST['STORAGE_VERIFY_CERT']
+    events_per_job = settings.GRIMOIRELAB_ARCHIVIST['EVENTS_PER_JOB']
+
+    if clear_tasks:
+        StorageTask.objects.all().delete()
+        click.echo("Removing old background tasks.")
+
+    current = StorageTask.objects.filter(burst=False).count()
+    if workers == current:
+        click.echo("Background tasks already created. Skipping.")
+        return
+
+    task_args = {
+        'storage_url': storage_url,
+        'storage_db_name': storage_db_name,
+        'storage_verify_certs': verify_certs,
+        'redis_group': 'archivist',
+        'limit': events_per_job
+    }
+    if workers > current:
+        for _ in range(workers - current):
+            schedule_task(
+                task_type=StorageTask.TASK_TYPE,
+                storage_type=storage_type,
+                task_args=task_args,
+                job_interval=1,
+                job_max_retries=10
+            )
+        click.echo(f"Created {workers} background tasks.")
+    elif workers < current:
+        tasks = StorageTask.objects.all()[workers:]
+        tasks.update(burst=True)

src/grimoirelab/core/runner/commands/run.py

@@ -357,3 +357,11 @@ def get_all_registered_task_models() -> Iterator[type[Task], type[Job]]:
         job classes.
     """
     return iter(GRIMOIRELAB_TASK_MODELS.values())
+
+
+def get_all_registered_task_names() -> list[str]:
+    """Return all registered task names.
+
+    :returns: a list with all registered task names.
+    """
+    return list(GRIMOIRELAB_TASK_MODELS.keys())

src/grimoirelab/core/scheduler/models.py

@@ -100,7 +100,7 @@ def cancel_task(task_uuid: str) -> None:
 
     jobs = job_class.objects.filter(task=task).all()
     for job in jobs:
-        job_rq = rq.job.Job.fetch(job.uuid, connection=django_rq.get_connection())
+        job_rq = rq.job.Job.fetch(job.uuid, connection=django_rq.get_connection(task.default_job_queue))
         job_rq.delete()
 
     task.delete()
@@ -125,7 +125,7 @@ def maintain_tasks() -> None:
         job_db = task.jobs.order_by('scheduled_at').first()
 
         try:
-            rq.job.Job.fetch(job_db.uuid, connection=django_rq.get_connection())
+            rq.job.Job.fetch(job_db.uuid, connection=django_rq.get_connection(task.default_job_queue))
             continue
         except rq.exceptions.NoSuchJobError:
             logger.debug(