[IMP] maintenance_service_http_monitoring : modify logic to avoid false-positive http alerts

maintenance_service_http_monitoring/models/service_instance.py

@@ -1,5 +1,5 @@
 import logging
-import time
+from datetime import timedelta
 
 from odoo import api, fields, models
 
@@ -10,8 +10,8 @@ except ImportError:
 
 _logger = logging.getLogger(__name__)
 
-HTTP_CHECK_TIMEOUT = 10  # seconds
-HTTP_RETRY_DELAY = 2  # seconds between pass 1 and pass 2
+HTTP_CHECK_TIMEOUT = 20  # seconds
+HTTP_KO_CONFIRMATION_DELAY = timedelta(minutes=5)
 
 
 class ServiceInstance(models.Model):
@@ -36,14 +36,20 @@ class ServiceInstance(models.Model):
         string="HTTP Maintenance Request",
         readonly=True,
     )
+    http_first_ko_at = fields.Datetime(
+        string="First HTTP KO Check (current streak)",
+        readonly=True,
+    )
 
     def check_http_status(self):
         """
         Perform HTTP check for each record and return the KO recordset.
 
         Writes last_http_status_code, last_http_check_date and http_status_ok on every
-        checked record. Does NOT create maintenance.request — that decision belongs to
-        the caller (cron) after optional retry logic.
+        checked record. Does NOT create maintenance.request — the cron only opens one
+        once the service has been continuously KO for at least
+        HTTP_KO_CONFIRMATION_DELAY, to avoid flagging transient outages (e.g. a short
+        server overload) as real incidents.
         """
         ko_records = self.browse()
         for rec in self:
@@ -63,13 +69,16 @@ class ServiceInstance(models.Model):
                 status_ok = status_code == 200
             except requests.exceptions.RequestException as e:
                 _logger.warning("HTTP check failed for %s: %s", rec.service_url, e)
-            rec.write(
-                {
-                    "last_http_status_code": status_code,
-                    "last_http_check_date": now,
-                    "http_status_ok": status_ok,
-                }
-            )
+            vals = {
+                "last_http_status_code": status_code,
+                "last_http_check_date": now,
+                "http_status_ok": status_ok,
+            }
+            if status_ok:
+                vals["http_first_ko_at"] = False
+            elif not rec.http_first_ko_at:
+                vals["http_first_ko_at"] = now
+            rec.write(vals)
             if not status_ok:
                 ko_records |= rec
         return ko_records
@@ -106,13 +115,12 @@ class ServiceInstance(models.Model):
     @api.model
     def cron_check_http_services(self):
         """
-        Check all active services with a URL, with one retry on failure.
+        Check all active services with a URL.
 
-        Pass 1: test every eligible service.
-        - Services that had an open request and are now OK are auto-resolved.
-        - Services still KO after pass 1 are retested after HTTP_RETRY_DELAY seconds.
-        maintenance.request is created only for services that fail both passes,
-        reducing noise from transient HTTP errors.
+        A service must be continuously KO for at least HTTP_KO_CONFIRMATION_DELAY
+        before a maintenance.request is created — this tolerates transient outages
+        (e.g. a temporary server overload) regardless of how often this cron runs.
+        Services that had an open request and are now OK are auto-resolved.
         """
         domain = [
             ("active", "=", True),
@@ -123,25 +131,22 @@ class ServiceInstance(models.Model):
             lambda s: not s.equipment_id.maintenance_mode
         )
 
-        # Snapshot services that currently have an open request before pass 1
+        # Snapshot services that currently have an open request before the check
         services_with_open_request = services.filtered(
             lambda s: s.http_maintenance_request
             and not s.http_maintenance_request.stage_id.done
         )
 
-        ko_after_pass1 = services.check_http_status()
+        ko_services = services.check_http_status()
 
-        # Auto-resolve services that recovered during pass 1
+        # Auto-resolve services that recovered
         recovered = services_with_open_request.filtered(lambda s: s.http_status_ok)
         if recovered:
             recovered._close_http_maintenance_request()
 
-        if not ko_after_pass1:
-            return
-
-        time.sleep(HTTP_RETRY_DELAY)
-
-        ko_confirmed = ko_after_pass1.check_http_status()
-
-        for service in ko_confirmed:
+        confirmed_ko = ko_services.filtered(
+            lambda s: s.last_http_check_date - s.http_first_ko_at
+            >= HTTP_KO_CONFIRMATION_DELAY
+        )
+        for service in confirmed_ko:
             service.equipment_id.create_http_maintenance_request(service)