-
Notifications
You must be signed in to change notification settings - Fork 652
Configurable settings
There are various things that you can configure using App Settings, which you can find on the Configure page in the Azure portal.
Note that in addition to using App Settings, you can specify those settings in your .deployment file. This is useful if you want the setting to be part of your repository.
Use this flag to add things at the end of the msbuild command line, such that it overrides any previous parts of the default command line.
e.g. to choose the Debug build configuration (default is Release) and apply a chained config transform, you could have:
[config]
SCM_BUILD_ARGS=-p:Configuration=Debug;PublishProfile=MyChainedTransform
As another example, to do a clean build, you can use:
[config]
SCM_BUILD_ARGS=-t:Clean;Compile
Kudu uses the azure site deploymentscript
command described here to generate a deployment script. By default, it figures out what parameters to pass by looking at the files in the repo to determine the project type (e.g. Node, ASP.NET, ...).
But in some cases, you may want to override that and take control of the command line, which you can do using this setting.
e.g. to force your repo to be treated as a plain web site (no build), you can use:
SCM_SCRIPT_GENERATOR_ARGS=--basic -p FolderToDeploy
Depending on the type of site you are deploying and the deployment method you are using, Kudu will executes build steps on your site during deployment, such as npm install
or dotnet build
/dotnet publish
. To disable this, such as when performing deployment of an already-built site, set:
SCM_DO_BUILD_DURING_DEPLOYMENT=false
This provides the same behavior as the SCM_SCRIPT_GENERATOR_ARGS=--basic
example above.
The default for zip deployments is false
, and for all other deployment methods is true
.
Every ZipDeploy, we will keep at most 5 recent zip files. You can keep more or less by adding below appSettings.
SCM_MAX_ZIP_PACKAGE_COUNT=N
See Customizing deployments for details. e.g.
COMMAND=deploy.cmd
Please see Deploying inplace and without repository for information on using the SCM_REPOSITORY_PATH
, SCM_NO_REPOSITORY
, PROJECT
and SCM_TARGET_PATH
flags.
By default, Kudu automatically update submodules before doing a deployment. To turn that off:
SCM_DISABLE_SUBMODULES=1
Note: this is support as an Azure App Setting, but not in the .deployment file.
By default, Kudu automatically creates a temporary app_offline.htm (if it is not already present) before doing a deployment. To turn that off:
SCM_CREATE_APP_OFFLINE=0
For large repos, you can make Kudu use a shallow clone when it clones your repo from GitHub or Bitbucket, which can save disk space. Shallow clones can be tricky, so make sure you understand what they are before using this. It is off by default. To turn it on:
SCM_USE_SHALLOW_CLONE=1
Note: this is support as an Azure App Setting, but not in the .deployment file.
By default, it gets touched. Sometimes, it's sub-optimal as it can causes an unnecessary restart. To avoid it, set:
SCM_TOUCH_WEBCONFIG_AFTER_DEPLOYMENT=0
See Post Deployment Action Hooks for details.
With this flag, a git push
updates the server repo, but does not trigger a deployment. Note that this only applies when git pushing directly to the Kudu git endpoint, and not in continuous integration scenarios like GitHub/Bitbucket.
SCM_DISABLE_DEPLOY_ON_PUSH=1
By default, kudu starts deployment as soon as git repository is updated. You can delay this deployment for X seconds, X being a random number between 0 and SCM_MAX_RANDOM_START_DELAY
SCM_MAX_RANDOM_START_DELAY=100
funcpack runs WebPack, turning large npm trees into single files, which can improve cold start greatly.
SCM_USE_FUNCPACK=1
The environment variable WEBSITE_HOSTNAME
contains the current host name of the slot. However it is by default slot sticky so that it will not cause a production slot restart during swap.
By default, it is set to 1, but you can get more tracing with higher values, up to 4. e.g.
SCM_TRACE_LEVEL=4
By default, when your build process launches some command, it's allowed to run for up to 60 seconds without producing any output. If that is not long enough, you can make it longer, e.g. to make it 10 minutes:
SCM_COMMAND_IDLE_TIMEOUT=600
Note that on Azure, there is a general idle request timeout that will cause clients to get disconnected after 230 seconds. However, the command will still continue running server-side after that.
When using the log streaming feature, by default it times out after 30 minutes of inactivity. To change it to 15 minutes (unit is seconds):
SCM_LOGSTREAM_TIMEOUT=900
Each site gets the site extensions feed from a configurable Url. If it is not set, the behavior is equivalent to
SCM_SITEEXTENSIONS_FEED_URL=https://www.nuget.org/api/v2/
For reference, the old feed was:
SCM_SITEEXTENSIONS_FEED_URL=http://www.siteextensions.net/api/v2/
The following settings must be set in the Azure App Settings, and cannot be overridden in the .deployment file (since they are not deployment settings)
This is often useful to workaround LIBGIT2SHARP hangs fetching repo (especially large one).
SCM_USE_LIBGIT2SHARP_REPOSITORY=0
Used the change the version of Node that is used by default
WEBSITE_NODE_DEFAULT_VERSION=0.10.5
Used the change the version of Npm that is used by default
WEBSITE_NPM_DEFAULT_VERSION=1.2.30
See this forum thread for details.
Important: this is only available for sites running in the Basic,Standard, and Premium mode.
WEBSITE_LOAD_USER_PROFILE=1
If you new up X509Certificate2 instance from a pfx file or cert encoded containing private key, you may run into CryptographicException: The system cannot find the file specified.. The reason is Windows stores private key as a file under a user profile directory. By default, Azure Web App (AppService) does not load user profile (avoid overhead for majority scenarios where it is not needed). Hence, the The system cannot find the file specified. issue. To work around, set the below appSetting to enable User Profile.
WEBSITE_LOAD_USER_PROFILE=1
This feature copies the site bits to the faster local drive before running them. Se this post for more info.
WEBSITE_LOCAL_CACHE_OPTION=Always
Also, the size of the cache can be changed. The default is 1GB (1000MB). The following app setting increases the local cache to 1.5GB (1500MB). You can increase the size up to 2 GB (2000 MB) per app.
WEBSITE_LOCAL_CACHE_SIZEINMB=1500
By default, Dynamic Cache is turned on in App Service and the WEBSITE_DYNAMIC_CACHE is 1.
Due to network file shared nature to allow multiple instance access, the dynamic cache improves performance by caching the recently accessed files locally on an instance. Cache is invalidated when file is modified. The cache location is %SYSTEMDRIVE%\local\DynamicCache
(same %SYSTEMDRIVE%\local
quota is applied).
Full content caching: caches both file content and directory/file metadata (timestamps, size, directory content):
WEBSITE_DYNAMIC_CACHE=1
Directory metadata caching: will not cache content of the files, only the directory/file metadata (timestamps, size, directory content). That results in much less local disk use:
WEBSITE_DYNAMIC_CACHE=2
This is the default behaviour for both v1 and v2 Azure Functions apps. HTTP Scaling
WEBSITE_HTTPSCALEV2_ENABLED=1
https://github.com/Azure/app-service-announcements/issues/84
WEBSITE_RUN_FROM_PACKAGE=https://davidebbostorage.blob.core.windows.net/content/SampleCoreMVCApp.zip?st=2018-02-13T09%3A48%3A00Z&se=2044-06-14T09%3A48%3A00Z&sp=rl&sv=2017-04-17&sr=b&sig=bNrVrEFzRHQB17GFJ7boEanetyJ9DGwBSV8OM3Mdh%2FM%3D
WEBSITE_PRIVATE_EXTENSIONS=0
By default the versions for site extensions are specific to each slot. This prevents unanticipated application behavior due to changing extension versions after a swap. If you want the extension versions to swap as well, create the application setting below on all slots.
WEBSITES_OVERRIDE_STICKY_EXTENSION_VERSIONS=1
See this post for details. First, the certs must be uploaded to the App Service Plan. Then App settings need to be set:
WEBSITE_LOAD_CERTIFICATES=*
or the specific thumbprints for just one
WEBSITE_LOAD_CERTIFICATES=ABCABCABCABCABCABCABCABCABCABCABCABCABCABC,DEFDEFDEFDEFDEFDEFDEFDEFDEFDEFDEFDEFDEFDEF
By default, the time zone is always UTC, but you can change it. You can get the list of valid values from this article (warning this other article shows slightly different strings which do not work!). You can also find the list in your registry under HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Time Zones
. If the string is not recognized, it falls back to UTC. The best way to test that it works is to type time
from Kudu console.
Examples:
WEBSITE_TIME_ZONE=Eastern Standard Time
WEBSITE_TIME_ZONE=AUS Eastern Standard Time
Note that there appears to be an issue that makes it not work when using DateTimeOffset.Now
. See this question for details.
By default, even on dedicated, the timeout is 20 minutes. But you can increase it:
WEBSITE_SCM_IDLE_TIMEOUT_IN_MINUTES=30
Note that this is not needed if you are also using the Always On feature.
It's an optimization that improves the cold start. In case it causes problems, it can be disabled:
WEBSITE_USE_PLACEHOLDER=0
Overlapped recycling makes it so that before the current instance on an app is shut down, a new instance starts. It can in some cases cause file locking issues, in which case you can try turning it off:
WEBSITE_DISABLE_OVERLAPPED_RECYCLING=1
The name (or relative path to the LogDirectory) of the file where internal errors are logged, for troubleshooting the listener:
DIAGNOSTICS_LASTRESORTFILE=logging-errors.txt
The settings file, relative to the web app root:
DIAGNOSTICS_LOGGINGSETTINGSFILE=..\diagnostics\settings.json
The log folder, relative to the web app root:
DIAGNOSTICS_TEXTTRACELOGDIRECTORY=..\..\LogFiles\Application
Maximum size of the log file (Default: 128 kb):
DIAGNOSTICS_TEXTTRACEMAXLOGFILESIZEBYTES=200000
Maximum size of the log folder (Default: 1 MB):
DIAGNOSTICS_TEXTTRACEMAXLOGFOLDERSIZEBYTES=2000000
Timeout in milliseconds to keep application logging on (Default is 43200000, which is 12 hours):
DIAGNOSTICS_TEXTTRACETURNOFFPERIOD=43200000
WEBSITE_WEBDEPLOY_USE_SCM=false
Note: this only applies if you don't set WEBSITE_WEBDEPLOY_USE_SCM=false
MSDEPLOY_RENAME_LOCKED_FILES=1
WEBSITE_DISABLE_SCM_SEPARATION=true
When separation enabled (the default), the main site and scm site run in different sandboxes. Some resulting behavior:
- With separation, when you stop the site, the scm site is still running, and you can continue to use git and msdeploy.
- With separation, the Main and scm sites each have their own local files. So you won't see the Main site's temp files from Kudu console.
Please be aware that turning off separation is considered to be a legacy mode that is no longer fully supported.
Note: this only applies if you don't set WEBSITE_WEBDEPLOY_USE_SCM=false
MSDEPLOY_RENAME_LOCKED_FILES=1
To ensure, REST api calls to update site and siteconfig are completely applied to all instances before returning. This is the default if using ARM template - to avoid race condition with subsequent ARM calls.
WEBSITE_ENABLE_SYNC_UPDATE_SITE=1
WEBSITE_START_SCM_ON_SITE_CREATION=true
Important: Set this app setting on all slots.
After slot swaps, the app may experience unexpected restarts. This is because after a swap, the hostname binding configuration goes out of sync, which by itself doesn't cause restarts. However, certain underlying storage events (such as storage volume failovers) may detect these discrepancies and force all worker processes to restart. To minimize these types of restarts, set the app setting below on all slots. Do not set this value if you are running a Windows Communication Foundation (WCF) application. If you are not using a WFC application there is no downside to creating this App Setting.
WEBSITE_ADD_SITENAME_BINDINGS_IN_APPHOST_CONFIG=1
See this
WEBSITE_SWAP_WARMUP_PING_PATH = The path to make the warm up request to. Set this to a URL path that begins with a slash as the value. For example, “/warmup.php”. The default value is /.
WEBSITE_SWAP_WARMUP_PING_STATUSES = Expected HTTP response codes for the warm-up operation. Set this to a comma-separated list of HTTP status codes. For example: “200,202” . If the returned status code is not in the list, the swap operation will not complete. By default, all response codes are valid.
The WEBSITE_MAX_DYNAMIC_APPLICATION_SCALE_OUT setting detailed here is no longer supported for controlling Functions scaling. The new setting is called functionAppScaleLimit and is detailed here .
See https://docs.microsoft.com/en-us/azure/app-service/containers/app-service-linux-faq for details on WEBSITES_ENABLE_APP_SERVICE_STORAGE
, WEBSITES_CONTAINER_START_TIME_LIMIT
, WEBSITES_PORT
and other settings.