Add ebs loss scenario

- Added scenario config
- Added new folder containing scenarion scripts
- Added scenario in run_kraken.py and kraken config

Resolves krkn-chaos#678

Signed-off-by: jtydlack <[email protected]>

config/config.yaml

@@ -46,6 +46,8 @@ kraken:
               - scenarios/kube/service_hijacking.yaml
         - syn_flood:
               - scenarios/kube/syn_flood.yaml
+        - aws:
+              - scenarios/openshift/detach_ebs_volumes.yaml
 
 cerberus:
     cerberus_enabled: False                                # Enable it when cerberus is previously installed

kraken/aws/detach_ebs_volumes.py

@@ -0,0 +1,152 @@
+import boto3
+import logging
+import re
+import time
+import yaml
+
+from krkn_lib.k8s import KrknKubernetes
+from krkn_lib.models.telemetry import ScenarioTelemetry
+from krkn_lib.telemetry.k8s import KrknTelemetryKubernetes
+
+
+def run(
+    scenarios_list: list[str],
+    wait_duration: int,
+    krkn_lib: KrknKubernetes,
+    telemetry: KrknTelemetryKubernetes,
+) -> (list[str], list[ScenarioTelemetry]):
+
+    logging.info("Detach ebs volumes scenario running...")
+    scenario_telemetries = list[ScenarioTelemetry]()
+    failed_post_scenarios = []
+    for scenario in scenarios_list:
+        scenario_telemetry = ScenarioTelemetry()
+        scenario_telemetry.scenario = scenario
+        scenario_telemetry.start_timestamp = time.time()
+        telemetry.set_parameters_base64(scenario_telemetry, scenario)
+
+        # Loading parameters from scenario config
+        with open(scenario) as stream:
+            scenario_config = yaml.safe_load(stream)
+
+        volume_ids = scenario_config["ebs_volume_id"]
+        volume_ids = re.split(r",+\s+|,+|\s+", volume_ids)
+        regions = scenario_config["region"]
+        # TODO support for multiple regions
+        # regions = re.split(r",+\s+|,+|\s+", regions)
+        aws_access_key_id = scenario_config["aws_access_key_id"]
+        aws_secret_access_key = scenario_config["aws_secret_access_key"]
+        chaos_duration = scenario_config["chaos_duration"]
+
+        # TODO implement detaching volumes based on tag and instance
+        volume_tag = scenario_config["ebs_volume_tag"]
+
+        # Get the EBS attachment details
+        ec2, ec2_client = get_ec2_session(
+            regions, aws_access_key_id, aws_secret_access_key
+        )
+        volume_details = get_ebs_volume_attachment_details(
+            volume_ids, ec2_client
+        )
+        logging.info("Obtaining attachment details...")
+        for volume in volume_details:
+            logging.info(
+                f"Volume {volume['VolumeId']} status: {volume['State']}"
+            )
+
+            # Try detach volume
+            detach_ebs_volume(volume, ec2_client, ec2, chaos_duration)
+
+        logging.info(
+            f"End of scenario {scenario}. "
+            f"Waiting for the specified duration {wait_duration}..."
+        )
+        time.sleep(wait_duration)
+
+        scenario_telemetry.exit_status = 0
+        scenario_telemetry.end_timestamp = time.time()
+        scenario_telemetries.append(scenario_telemetry)
+        logging.info(f"Scenario {scenario} successfully finished")
+
+    return failed_post_scenarios, scenario_telemetries
+
+
+def fail(
+    scenario_telemetry: ScenarioTelemetry,
+    scenario_telemetries: list[ScenarioTelemetry],
+):
+    scenario_telemetry.exit_status = 1
+    scenario_telemetry.end_timestamp = time.time()
+    scenario_telemetries.append(scenario_telemetry)
+
+
+def get_ebs_volume_attachment_details(volume_ids: list, ec2_client):
+    response = ec2_client.describe_volumes(VolumeIds=volume_ids)
+    volumes_details = response["Volumes"]
+    return volumes_details
+
+
+def get_ebs_volume_state(volume_id: str, ec2_resource):
+    volume = ec2_resource.Volume(volume_id)
+    state = volume.state
+    return state
+
+
+def detach_ebs_volume(volume: dict, ec2_client, ec2_resource, duration: int):
+    if volume["State"] == "in-use":
+        logging.info(f"Detaching volume {volume['VolumeId']}...")
+        ec2_client.detach_volume(VolumeId=volume['VolumeId'])
+        if check_attachment_state(volume, ec2_resource, "available") == 1:
+            return
+        logging.info(f"Volume {volume['VolumeId']} successfully detached")
+        logging.info("Waiting for chaos duration...")
+        time.sleep(duration)
+
+        # Attach volume back
+        attach_ebs_volume(volume, ec2_client)
+        if check_attachment_state(volume, ec2_resource, "in-use") == 1:
+            return
+        logging.info(f"Volume {volume['VolumeId']} successfully attached")
+
+
+def attach_ebs_volume(volume: dict, ec2_client):
+    for attachment in volume["Attachments"]:
+        ec2_client.attach_volume(
+            InstanceId=attachment["InstanceId"],
+            Device=attachment["Device"],
+            VolumeId=volume["VolumeId"],
+        )
+
+
+def get_ec2_session(
+    region: str, aws_access_key_id: str, aws_secret_access_key: str
+):
+    ec2 = boto3.resource(
+        "ec2",
+        region_name=region,
+        aws_access_key_id=aws_access_key_id,
+        aws_secret_access_key=aws_secret_access_key,
+    )
+    ec2_client = boto3.client(
+        "ec2",
+        region_name=region,
+        aws_access_key_id=aws_access_key_id,
+        aws_secret_access_key=aws_secret_access_key,
+    )
+    return ec2, ec2_client
+
+
+def check_attachment_state(volume, ec2, desired_state: str):
+    time.sleep(5)
+    state = get_ebs_volume_state(volume["VolumeId"], ec2)
+    for i in range(5):
+        if state == desired_state:
+            return 0
+        logging.debug(f"Volume in undesired state {state}...")
+        time.sleep(3)
+    else:
+        logging.error(
+            f"Something went wrong, last volume {volume['VolumeId']} "
+            f"state was {state}, desired state {desired_state}"
+        )
+        return 1

run_kraken.py

@@ -30,6 +30,7 @@
 import kraken.arcaflow_plugin as arcaflow_plugin
 import kraken.prometheus as prometheus_plugin
 import kraken.service_hijacking.service_hijacking as service_hijacking_plugin
+import kraken.aws.detach_ebs_volumes as detach_ebs_volumes
 import server as server
 from kraken import plugins, syn_flood
 from krkn_lib.k8s import KrknKubernetes
@@ -44,6 +45,7 @@
 
 report_file = ""
 
+
 # Main function
 def main(cfg) -> int:
     # Start kraken
@@ -464,6 +466,10 @@ def main(cfg) -> int:
                                                                                         telemetry_request_id
                                                                                         )
                             chaos_telemetry.scenarios.extend(scenario_telemetries)
+                        elif scenario_type == "aws":
+                            logging.info("Running AWS Chaos")
+                            failed_post_scenarios, scenario_telemetries = detach_ebs_volumes.run(scenarios_list, wait_duration, kubecli, telemetry_k8s)
+                            # chaos_telemetry.scenarios.extend(scenario_telemetries)
 
                         # Check for critical alerts when enabled
                         post_critical_alerts = 0
@@ -481,7 +487,6 @@ def main(cfg) -> int:
                                 logging.error("Post chaos critical alerts firing please check, exiting")
                                 break
 
-
             iteration += 1
             logging.info("")
 

scenarios/openshift/detach_ebs_volumes.yaml

@@ -0,0 +1,9 @@
+aws_access_key_id:
+aws_secret_access_key:
+chaos_duration: 30
+
+region:
+ebs_volume_id:
+ebs_volume_tag:
+node_name:
+