zhangyiss/the-littlest-jupyterhub
1
0
Fork 0
mirror of https://github.com/jupyterhub/the-littlest-jupyterhub.git synced 2025-12-18 21:54:05 +08:00
Code Issues Packages Projects Releases Wiki Activity

Addressed docs feedback

Browse Source
This commit is contained in:
GeorgianaElena
2019-06-14 14:18:39 +03:00
parent 31d92199e1
commit 346701502d
1 changed files with 97 additions and 69 deletions
Download Patch File Download Diff File Expand all files Collapse all files

166
docs/topic/idle-culler.rst
View File

@@ -4,9 +4,102 @@
Culling idle notebook servers
=============================
The idle culler is a hub-managed service that automatically shuts down idle
single-user notebook servers in order to free up resources. After culling, any
in-memory data will be lost.
The idle culler automatically shuts down user notebook servers when they have
not been used for a certain time period, in order to reduce the total resource
usage on your JupyterHub.
JupyterHub pings the user's notebook server at certain time intervals. If no response
is received from the server during this checks and the timeout expires, the server is
considered to be *inactive (idle)* and will be culled.
Default settings
================
By default, JupyterHub will ping the user notebook servers every 60s to check their
status. Every server found to be idle for more than 10 minutes will be culled.
.. code-block:: python
services.cull.every = 60
services.cull.timeout = 600
Because the servers don't have a maximum age set, an active server will not be shut down
regardless of how long it has been up and running.
.. code-block:: python
services.cull.max_age = 0
If after the culling process, there are users with no active notebook servers, by default,
the users will not be culled alongside their notebooks and will continue to exist.
.. code-block:: python
services.cull.users = False
Configuring the idle culler
===========================
The available configuration options are:
Idle timeout
------------
The idle timeout is the maximum time (in seconds) a server can be inactive before it
will be culled. The timeout can be configured using:
.. code-block:: bash
sudo tljh-config set services.cull.timeout <max-idle-sec-before-server-is-culled>
sudo tljh-config reload
Idle check interval
-------------------
The idle check interval represents how frequent (in seconds) the Hub will
check if there are any idle servers to cull. It can be configured using:
.. code-block:: bash
sudo tljh-config set services.cull.every <number-of-sec-this-check-is-done>
sudo tljh-config reload
Maximum age
-----------
The maximum age sets the time (in seconds) a server should be running.
The servers that exceed the maximum age, will be culled even if they are active.
A maximum age of 0, will deactivate this option.
The maximum age can be configured using:
.. code-block:: bash
sudo tljh-config set services.cull.max_age <server-max-age>
sudo tljh-config reload
User culling
------------
In addition to servers, it is also possible to cull the users. This is usually
suited for temporary-user cases such as *tmpnb*.
User culling can be activated using the following command:
.. code-block:: bash
sudo tljh-config set services.cull.users True
sudo tljh-config reload
Concurrency
-----------
Deleting a lot of users at the same time can slow down the Hub.
The number of concurrent requests made to the Hub can be configured using:
.. code-block:: bash
sudo tljh-config set services.cull.concurrency <number-of-concurrent-hub-requests>
sudo tljh-config reload
Because TLJH it's used for a small number of users, the cases that may require to
modify the concurrency limit should be rare.
Disabling the idle culler
@@ -18,69 +111,4 @@ command:
.. code-block:: bash
sudo tljh-config set services.cull.enabled False
Configuring the idle culler
===========================
By **default**, JupyterHub will:
* Run the culling process every minute.
* Cull any user servers that have been inactive for more than 10 minutes.
The configuration options available are:
Idle timeout
------------
The idle timeout (in seconds) can be configured using:
.. code-block:: bash
sudo tljh-config set services.cull.timeout <max-idle-sec-before-server-is-culled>
*By default services.cull.timeout = 600*
Idle check interval
-------------------
The interval (in seconds) for checking for idle servers to cull can be configured using:
.. code-block:: bash
sudo tljh-config set services.cull.every <number-of-sec-this-check-is-done>
*By default services.cull.every = 60*
Maximum age
-----------
The maximum age (in seconds) of servers that should be culled even if they are active
can be configured using:
.. code-block:: bash
sudo tljh-config set services.cull.max_age <server-max-age>
*By default services.cull.max_age = 0*
User culling
------------
In addition to servers, the users will also be culled if the following command is used:
.. code-block:: bash
sudo tljh-config set services.cull.users True
*By default services.cull.users = False*
Concurrency
-----------
The number of concurrent requests made to the Hub ca be configured using:
.. code-block:: bash
sudo tljh-config set services.cull.concurrency <number-of-concurrent-hub-requests>
*By default services.cull.concurrency = 5*
sudo tljh-config reload