Feature 10981: added Mongo accesses needed for NGSA 84/12984/11
authoraguilard <e.dah.tid@telefonica.com>
Thu, 16 Feb 2023 17:24:35 +0000 (17:24 +0000)
committeraguilard <e.dah.tid@telefonica.com>
Tue, 9 May 2023 06:56:55 +0000 (06:56 +0000)
Change-Id: If3942d060f468382c7796a7e610bce9b21ab93fc
Signed-off-by: aguilard <e.dah.tid@telefonica.com>
osm_lcm/ns.py

index 0fce107..83869bc 100644 (file)
@@ -2327,6 +2327,212 @@ class NsLcm(LcmBase):
                 vnfr["vim-account-id"] = pla_vnf["vimAccountId"]
         return modified
 
+    def _gather_vnfr_healing_alerts(self, vnfr, vnfd):
+        alerts = []
+        nsr_id = vnfr["nsr-id-ref"]
+        df = vnfd.get("df", [{}])[0]
+        # Checking for auto-healing configuration
+        if "healing-aspect" in df:
+            healing_aspects = df["healing-aspect"]
+            for healing in healing_aspects:
+                for healing_policy in healing.get("healing-policy", ()):
+                    vdu_id = healing_policy["vdu-id"]
+                    vdur = next(
+                        (vdur for vdur in vnfr["vdur"] if vdu_id == vdur["vdu-id-ref"]),
+                        {},
+                    )
+                    if not vdur:
+                        continue
+                    metric_name = "vm_status"
+                    vdu_name = vdur.get("name")
+                    vnf_member_index = vnfr["member-vnf-index-ref"]
+                    uuid = str(uuid4())
+                    name = f"healing_{uuid}"
+                    action = healing_policy
+                    # action_on_recovery = healing.get("action-on-recovery")
+                    # cooldown_time = healing.get("cooldown-time")
+                    # day1 = healing.get("day1")
+                    alert = {
+                        "uuid": uuid,
+                        "name": name,
+                        "metric": metric_name,
+                        "tags": {
+                            "ns_id": nsr_id,
+                            "vnf_member_index": vnf_member_index,
+                            "vdu_name": vdu_name,
+                        },
+                        "alarm_status": "ok",
+                        "action_type": "healing",
+                        "action": action,
+                    }
+                    alerts.append(alert)
+        return alerts
+
+    def _gather_vnfr_scaling_alerts(self, vnfr, vnfd):
+        alerts = []
+        nsr_id = vnfr["nsr-id-ref"]
+        df = vnfd.get("df", [{}])[0]
+        # Checking for auto-scaling configuration
+        if "scaling-aspect" in df:
+            rel_operation_types = {
+                "GE": ">=",
+                "LE": "<=",
+                "GT": ">",
+                "LT": "<",
+                "EQ": "==",
+                "NE": "!=",
+            }
+            scaling_aspects = df["scaling-aspect"]
+            all_vnfd_monitoring_params = {}
+            for ivld in vnfd.get("int-virtual-link-desc", ()):
+                for mp in ivld.get("monitoring-parameters", ()):
+                    all_vnfd_monitoring_params[mp.get("id")] = mp
+            for vdu in vnfd.get("vdu", ()):
+                for mp in vdu.get("monitoring-parameter", ()):
+                    all_vnfd_monitoring_params[mp.get("id")] = mp
+            for df in vnfd.get("df", ()):
+                for mp in df.get("monitoring-parameter", ()):
+                    all_vnfd_monitoring_params[mp.get("id")] = mp
+            for scaling_aspect in scaling_aspects:
+                scaling_group_name = scaling_aspect.get("name", "")
+                # Get monitored VDUs
+                all_monitored_vdus = set()
+                for delta in scaling_aspect.get("aspect-delta-details", {}).get(
+                    "deltas", ()
+                ):
+                    for vdu_delta in delta.get("vdu-delta", ()):
+                        all_monitored_vdus.add(vdu_delta.get("id"))
+                monitored_vdurs = list(
+                    filter(
+                        lambda vdur: vdur["vdu-id-ref"] in all_monitored_vdus,
+                        vnfr["vdur"],
+                    )
+                )
+                if not monitored_vdurs:
+                    self.logger.error(
+                        "Scaling criteria is referring to a vnf-monitoring-param that does not contain a reference to a vdu or vnf metric"
+                    )
+                    continue
+                for scaling_policy in scaling_aspect.get("scaling-policy", ()):
+                    if scaling_policy["scaling-type"] != "automatic":
+                        continue
+                    threshold_time = scaling_policy.get("threshold-time", "1")
+                    cooldown_time = scaling_policy.get("cooldown-time", "0")
+                    for scaling_criteria in scaling_policy["scaling-criteria"]:
+                        monitoring_param_ref = scaling_criteria.get(
+                            "vnf-monitoring-param-ref"
+                        )
+                        vnf_monitoring_param = all_vnfd_monitoring_params[
+                            monitoring_param_ref
+                        ]
+                        for vdur in monitored_vdurs:
+                            vdu_id = vdur["vdu-id-ref"]
+                            metric_name = vnf_monitoring_param.get("performance-metric")
+                            vnf_member_index = vnfr["member-vnf-index-ref"]
+                            scalein_threshold = scaling_criteria.get(
+                                "scale-in-threshold"
+                            )
+                            scaleout_threshold = scaling_criteria.get(
+                                "scale-out-threshold"
+                            )
+                            # Looking for min/max-number-of-instances
+                            instances_min_number = 1
+                            instances_max_number = 1
+                            vdu_profile = df["vdu-profile"]
+                            if vdu_profile:
+                                profile = next(
+                                    item for item in vdu_profile if item["id"] == vdu_id
+                                )
+                                instances_min_number = profile.get(
+                                    "min-number-of-instances", 1
+                                )
+                                instances_max_number = profile.get(
+                                    "max-number-of-instances", 1
+                                )
+
+                            if scalein_threshold:
+                                uuid = str(uuid4())
+                                name = f"scalein_{uuid}"
+                                operation = scaling_criteria[
+                                    "scale-in-relational-operation"
+                                ]
+                                rel_operator = rel_operation_types.get(operation, "<=")
+                                metric_selector = f'{metric_name}{{ns_id="{nsr_id}", vnf_member_index="{vnf_member_index}", vdu_id="{vdu_id}"}}'
+                                expression = f"(count ({metric_selector}) > {instances_min_number}) and (avg({metric_selector}) {rel_operator} {scalein_threshold})"
+                                labels = {
+                                    "ns_id": nsr_id,
+                                    "vnf_member_index": vnf_member_index,
+                                    "vdu_id": vdu_id,
+                                }
+                                prom_cfg = {
+                                    "alert": name,
+                                    "expr": expression,
+                                    "for": str(threshold_time) + "m",
+                                    "labels": labels,
+                                }
+                                action = scaling_policy
+                                action = {
+                                    "scaling-group": scaling_group_name,
+                                    "cooldown-time": cooldown_time,
+                                }
+                                alert = {
+                                    "uuid": uuid,
+                                    "name": name,
+                                    "metric": metric_name,
+                                    "tags": {
+                                        "ns_id": nsr_id,
+                                        "vnf_member_index": vnf_member_index,
+                                        "vdu_id": vdu_id,
+                                    },
+                                    "alarm_status": "ok",
+                                    "action_type": "scale_in",
+                                    "action": action,
+                                    "prometheus_config": prom_cfg,
+                                }
+                                alerts.append(alert)
+
+                            if scaleout_threshold:
+                                uuid = str(uuid4())
+                                name = f"scaleout_{uuid}"
+                                operation = scaling_criteria[
+                                    "scale-out-relational-operation"
+                                ]
+                                rel_operator = rel_operation_types.get(operation, "<=")
+                                metric_selector = f'{metric_name}{{ns_id="{nsr_id}", vnf_member_index="{vnf_member_index}", vdu_id="{vdu_id}"}}'
+                                expression = f"(count ({metric_selector}) < {instances_max_number}) and (avg({metric_selector}) {rel_operator} {scaleout_threshold})"
+                                labels = {
+                                    "ns_id": nsr_id,
+                                    "vnf_member_index": vnf_member_index,
+                                    "vdu_id": vdu_id,
+                                }
+                                prom_cfg = {
+                                    "alert": name,
+                                    "expr": expression,
+                                    "for": str(threshold_time) + "m",
+                                    "labels": labels,
+                                }
+                                action = scaling_policy
+                                action = {
+                                    "scaling-group": scaling_group_name,
+                                    "cooldown-time": cooldown_time,
+                                }
+                                alert = {
+                                    "uuid": uuid,
+                                    "name": name,
+                                    "metric": metric_name,
+                                    "tags": {
+                                        "ns_id": nsr_id,
+                                        "vnf_member_index": vnf_member_index,
+                                        "vdu_id": vdu_id,
+                                    },
+                                    "alarm_status": "ok",
+                                    "action_type": "scale_out",
+                                    "action": action,
+                                    "prometheus_config": prom_cfg,
+                                }
+                                alerts.append(alert)
+        return alerts
+
     def update_nsrs_with_pla_result(self, params):
         try:
             nslcmop_id = deep_get(params, ("placement", "nslcmopId"))
@@ -2801,6 +3007,22 @@ class NsLcm(LcmBase):
                 db_nsr_update["detailed-status"] = "Done"
                 db_nslcmop_update["detailed-status"] = "Done"
                 nslcmop_operation_state = "COMPLETED"
+                # Gather auto-healing and auto-scaling alerts for each vnfr
+                healing_alerts = []
+                scaling_alerts = []
+                for vnfr in self.db.get_list("vnfrs", {"nsr-id-ref": nsr_id}):
+                    vnfd = next(
+                        (sub for sub in db_vnfds if sub["_id"] == vnfr["vnfd-id"]), None
+                    )
+                    healing_alerts = self._gather_vnfr_healing_alerts(vnfr, vnfd)
+                    for alert in healing_alerts:
+                        self.logger.info(f"Storing healing alert in MongoDB: {alert}")
+                        self.db.create("alerts", alert)
+
+                    scaling_alerts = self._gather_vnfr_scaling_alerts(vnfr, vnfd)
+                    for alert in scaling_alerts:
+                        self.logger.info(f"Storing scaling alert in MongoDB: {alert}")
+                        self.db.create("alerts", alert)
 
             if db_nsr:
                 self._write_ns_status(
@@ -4523,6 +4745,8 @@ class NsLcm(LcmBase):
                     self.logger.error(
                         logging_text + "kafka_write notification Exception {}".format(e)
                     )
+                self.logger.debug(f"Deleting alerts: ns_id={nsr_id}")
+                self.db.del_list("alerts", {"tags.ns_id": nsr_id})
 
             self.logger.debug(logging_text + "Exit")
             self.lcm_tasks.remove("ns", nsr_id, nslcmop_id, "ns_terminate")