diff --git a/README.md b/README.md
index 3e04e048..c8b47ce7 100644
--- a/README.md
+++ b/README.md
@@ -17,7 +17,7 @@ sudo snap install charmed-openstack-upgrader
 
 An alias `cou` will be automatically enabled upon successful installation.
 
-Run `cou -h` to learn about the available commands: 
+Run `cou -h` to learn about the available commands:
 
 ```bash
 Usage: cou [options] <command>
@@ -43,6 +43,8 @@ Commands:
 - `COU_TIMEOUT` - define timeout for cou retry policy. Default value is 10 seconds.
 - `COU_MODEL_RETRIES` - define how many times to retry the connection to Juju model before giving up. Default value is 5 times.
 - `COU_MODEL_RETRY_BACKOFF` - define number of seconds to increase the wait between connection to the Juju model retry attempts. Default value is 2 seconds.
+- `COU_STANDARD_IDLE_TIMEOUT` - how long COU will wait for an application to settle to active/idle and declare the upgrade complete. The default value is 300 seconds.
+- `COU_LONG_IDLE_TIMEOUT` - a longer version of COU_STANDARD_IDLE_TIMEOUT for applications that are known to need more time than usual to upgrade like such as Keystone and Octavia. The default value is 1800 seconds.
 
 ## Supported Upgrade Paths
 
@@ -57,4 +59,4 @@ upgrades.
 
 # License
 Charmed OpenStack Upgrader is a free software, distributed under the Apache-2.0 license. Refer to the
-[LICENSE](https://github.com/canonical/snap-tempest/blob/main/LICENSE) file for details.
\ No newline at end of file
+[LICENSE](https://github.com/canonical/snap-tempest/blob/main/LICENSE) file for details.
diff --git a/cou/apps/auxiliary.py b/cou/apps/auxiliary.py
index 8c0f9290..5a9a57cb 100644
--- a/cou/apps/auxiliary.py
+++ b/cou/apps/auxiliary.py
@@ -15,7 +15,7 @@
 import logging
 from typing import Optional
 
-from cou.apps.base import OpenStackApplication
+from cou.apps.base import LONG_IDLE_TIMEOUT, OpenStackApplication
 from cou.apps.factory import AppFactory
 from cou.exceptions import ApplicationError
 from cou.steps import PreUpgradeStep
@@ -128,7 +128,7 @@ class RabbitMQServer(OpenStackAuxiliaryApplication):
     RabbitMQ must wait for the entire model to be idle before declaring the upgrade complete.
     """
 
-    wait_timeout = 30 * 60  # 30 min
+    wait_timeout = LONG_IDLE_TIMEOUT
     wait_for_model = True
 
 
@@ -136,7 +136,7 @@ class RabbitMQServer(OpenStackAuxiliaryApplication):
 class CephMonApplication(OpenStackAuxiliaryApplication):
     """Application for Ceph Monitor charm."""
 
-    wait_timeout = 30 * 60  # 30 min
+    wait_timeout = LONG_IDLE_TIMEOUT
     wait_for_model = True
 
     def pre_upgrade_steps(self, target: OpenStackRelease) -> list[PreUpgradeStep]:
@@ -189,7 +189,7 @@ class MysqlInnodbClusterApplication(OpenStackAuxiliaryApplication):
     # NOTE(agileshaw): holding 'mysql-server-core-8.0' package prevents undesired
     # mysqld processes from restarting, which lead to outages
     packages_to_hold: Optional[list] = ["mysql-server-core-8.0"]
-    wait_timeout = 30 * 60  # 30 min
+    wait_timeout = LONG_IDLE_TIMEOUT
 
 
 # NOTE (gabrielcocenza): Although CephOSD class is empty now, it will be
diff --git a/cou/apps/base.py b/cou/apps/base.py
index f59629bd..f3f8f5fb 100644
--- a/cou/apps/base.py
+++ b/cou/apps/base.py
@@ -16,6 +16,7 @@
 from __future__ import annotations
 
 import logging
+import os
 from collections import defaultdict
 from dataclasses import dataclass, field
 from io import StringIO
@@ -46,7 +47,10 @@
 
 logger = logging.getLogger(__name__)
 
-DEFAULT_WAITING_TIMEOUT = 5 * 60  # 5 min
+STANDARD_IDLE_TIMEOUT: int = int(
+    os.environ.get("COU_STANDARD_IDLE_TIMEOUT", 5 * 60)
+)  # default of 5 min
+LONG_IDLE_TIMEOUT: int = int(os.environ.get("COU_LONG_IDLE_TIMEOUT", 30 * 60))  # default of 30 min
 
 
 @dataclass
@@ -93,7 +97,7 @@ class OpenStackApplication:
     charm: str
     units: list[ApplicationUnit] = field(default_factory=lambda: [])
     packages_to_hold: Optional[list] = field(default=None, init=False)
-    wait_timeout: int = field(default=DEFAULT_WAITING_TIMEOUT, init=False)
+    wait_timeout: int = field(default=STANDARD_IDLE_TIMEOUT, init=False)
     wait_for_model: bool = field(default=False, init=False)  # waiting only for application itself
 
     def __post_init__(self) -> None:
diff --git a/cou/apps/core.py b/cou/apps/core.py
index 1acf9e71..c153f359 100644
--- a/cou/apps/core.py
+++ b/cou/apps/core.py
@@ -15,7 +15,7 @@
 """Core application class."""
 import logging
 
-from cou.apps.base import OpenStackApplication
+from cou.apps.base import LONG_IDLE_TIMEOUT, OpenStackApplication
 from cou.apps.factory import AppFactory
 
 logger = logging.getLogger(__name__)
@@ -28,7 +28,7 @@ class Keystone(OpenStackApplication):
     Keystone must wait for the entire model to be idle before declaring the upgrade complete.
     """
 
-    wait_timeout = 30 * 60  # 30 min
+    wait_timeout = LONG_IDLE_TIMEOUT
     wait_for_model = True
 
 
@@ -39,4 +39,4 @@ class Octavia(OpenStackApplication):
     Octavia required more time to settle before COU can continue.
     """
 
-    wait_timeout = 30 * 60  # 30 min
+    wait_timeout = LONG_IDLE_TIMEOUT
diff --git a/docs/how-to/different-model.rst b/docs/how-to/different-model.rst
index ac05386f..29a9be63 100644
--- a/docs/how-to/different-model.rst
+++ b/docs/how-to/different-model.rst
@@ -29,5 +29,5 @@ configuration files.
 
 .. LINKS
 .. _python-libjuju: https://github.com/juju/python-libjuju
-.. _Juju environment variables: https://juju.is/docs/juju/environment-variables#heading--jujudata
+.. _Juju environment variables: https://juju.is/docs/juju/environment-variables
 .. _JUJU_DATA: https://juju.is/docs/juju/environment-variables#heading--jujudata
diff --git a/docs/reference/environment-variables.rst b/docs/reference/environment-variables.rst
index 9dba5d55..7601dc7d 100644
--- a/docs/reference/environment-variables.rst
+++ b/docs/reference/environment-variables.rst
@@ -6,3 +6,5 @@ Environment Variables
 * **COU_TIMEOUT** - define timeout for **COU** retry policy. Default value is 10 seconds.
 * **COU_MODEL_RETRIES** - define how many times to retry the connection to Juju model before giving up. Default value is 5 times.
 * **COU_MODEL_RETRY_BACKOFF** - define number of seconds to increase the wait between connection to the Juju model retry attempts. Default value is 2 seconds.
+* **COU_STANDARD_IDLE_TIMEOUT** - how long COU will wait for an application to settle to active/idle and declare the upgrade complete. The default value is 300 seconds.
+* **COU_LONG_IDLE_TIMEOUT** - a longer version of COU_STANDARD_IDLE_TIMEOUT for applications that are known to need more time than usual to upgrade like such as Keystone and Octavia. The default value is 1800 seconds.