Python package configuration options

The following list includes all configuration options with the name of the environment variable and the name of the key in the configuration file.

For more information on how to configure AppSignal with a configuration file or system environment variables, see our Configuration documentation.

Available options

• Required options
  • active
  • environment
  • name
  • push_api_key
• Options
  • bind_address
  • ca_file_path
  • collector_endpoint
  • cpu_count
  • disable_default_instrumentations
  • dns_servers
  • enable_host_metrics
  • enable_minutely_probes
  • enable_nginx_metrics
  • enable_statsd
  • endpoint
  • files_world_accessible
  • filter_attributes
  • filter_function_parameters
  • filter_parameters
  • filter_request_payload
  • filter_request_query_parameters
  • filter_session_data
  • host_role
  • hostname
  • http_proxy
  • ignore_actions
  • ignore_errors
  • ignore_namespaces
  • log
  • log_level
  • log_path
  • nginx_port
  • opentelemetry_port
  • request_headers
  • response_headers
  • revision
  • running_in_container
  • send_environment_metadata
  • send_function_parameters
  • send_params
  • send_request_payload
  • send_request_query_parameters
  • send_session_data
  • service_name
  • statsd_port
  • working_directory_path

active

Description

Configure AppSignal to be active or not for a given environment. If AppSignal is not active, no data is reported for the app in the specified environment.

environment

Description

The environment of the app to be reported to AppSignal.

name

Description

Name of your application as it should be displayed on AppSignal.com.

push_api_key

Description

The organization-level authentication key to authenticate with our Push API.

Read more about the AppSignal Push API key.

bind_address

Description

A valid IPv4 address the AppSignal agent uses as a binding for its TCP and UDP servers. Use a specific address if you only want the agent to listen to requests made to that address. Set this option to 0.0.0.0 to allow to receive requests from hosts using any IP address. By default it only listens to requests made on the same host. This option is applied to all the agent servers (StatsD, OpenTelemetry and NGINX).

ca_file_path

Description

Use this option to point to another certificate file if there's a problem connecting to our API.

collector_endpoint

Description

The address of an externally deployed AppSignal Collector. When set, the AppSignal agent bundled with the integration will not be used, and OpenTelemetry data will be sent to the collector at this address instead.

cpu_count

Description

The available CPU capacity of the host, in number of CPUs. This is used to calculate the CPU usage percentage in the host metrics. If not set, the agent will attempt to automatically detect this from cgroups.

The number of CPUs can be a fraction, e.g. 0.5.

disable_default_instrumentations

Description

Allows you to disable the automatic instrumentation of default integrations:

• aiopg
• asyncpg
• celery
• django
• flask
• mysql
• mysqlclient
• psycopg
• psycopg2
• pymysql
• redis
• jinja2
• requests
• sqlalchemy
• sqlite3
• logging

Note that instrumentations are only enabled if the corresponding OpenTelemetry instrumentation library is installed.

To disable all automatic instrumentations, set the value to True.

dns_servers

Description

Configure DNS servers for the AppSignal agent to use.

# In the host environment
export APPSIGNAL_DNS_SERVERS="8.8.8.8,8.8.4.4"

If you're affected by our DNS timeouts, try setting a DNS server manually using this option that doesn't use more than 4 dots in the server name.

• Acceptable values: 8.8.8.8, my.custom.local.server.
• Not acceptable values: foo, my.awesome.custom.local.dns.server.

If the DNS server cannot be reached the agent will fall back on the host's DNS configuration and output a message in the appsignal.log file: A problem occurred while setting DNS servers.

enable_host_metrics

Description

Set this option to false to disable host metrics collection.

On Heroku and Dokku host metrics are disabled by default. This is done because these systems will report inaccurate metrics from within the containers. Host metrics collection on these systems cannot be enabled. For Heroku, use the Heroku log drain instead.

enable_minutely_probes

Description

Enables the minutely probes system.

enable_nginx_metrics

Description

Set to true to enable the NGINX metrics server. See the NGINX metrics documentation for details.

When enabled, the AppSignal agent will listen to a localhost-bound server on port 27649. If you're running several AppSignal-instrumented applications in the same server, this configuration option can only be enabled in one of them.

enable_statsd

Description

Enables the StatsD server in the AppSignal agent.

When enabled, the AppSignal agent will listen to a localhost-bound server on port 8125. If you're running several AppSignal-instrumented applications in the same server, this configuration option can only be enabled in one of them.

endpoint

Description

Configure the endpoint to send data to AppSignal. This setting will not have to be changed.

files_world_accessible

Description

If this is set to true the AppSignal working directory that is created is accessible for all users (Unix permissions 0666). This is often necessary because processes for the same app run under a different user. Set to false to disable this behaviour (Unix permissions 0644).

filter_attributes

Description

List of OpenTelemetry span attributes that should be ignored using AppSignal filtering. Their values will be replaced with [FILTERED] when transmitted to AppSignal.

Read more about parameter filtering.

filter_function_parameters

Description

List of function parameter attributes that should be ignored using AppSignal filtering. Their values will be replaced with [FILTERED] when transmitted to AppSignal.

Read more about parameter filtering.

filter_parameters

Description

List of parameter keys that should be ignored using AppSignal filtering. Their values will be replaced with [FILTERED] when transmitted to AppSignal.

Read more about parameter filtering.

filter_request_payload

Description

List of request payload attributes that should be ignored using AppSignal filtering. Their values will be replaced with [FILTERED] when transmitted to AppSignal.

Read more about parameter filtering.

filter_request_query_parameters

Description

List of request query parameter attributes that should be ignored using AppSignal filtering. Their values will be replaced with [FILTERED] when transmitted to AppSignal.

Read more about parameter filtering.

filter_session_data

Description

List of session data keys that should be ignored using AppSignal filtering. Their values will be replaced with [FILTERED] when transmitted to AppSignal.

Read more about session data filtering.

host_role

Description

Group hosts by role and generate metrics based on this role. One such metric is the reporting_hosts counter metric. A good role indicates what the main role of the server is, like "webserver", "processor", "api", "database", "loadbalancer", etc.

hostname

Description

This overrides the server's hostname. Useful for when you're unable to set a custom hostname or when a nondescript id is generated for you on hosting services.

http_proxy

Description

If you require the agent to connect to the Internet via a proxy set the complete proxy URL in this configuration key.

ignore_actions

Description

With this config option you can specify a list of actions that will be ignored by AppSignal. Everything that happens including exceptions will not be transmitted to AppSignal. This can be useful to ignore health check endpoints or other actions that you don't want to monitor.

Read more about ignoring actions.

ignore_errors

Description

List of error classes that will be ignored. Any exception raised with this error class will not be transmitted to AppSignal.

Read more about ignoring errors.

ignore_namespaces

Description

List of namespaces that will be ignored. Any error raised or slow request that occurs in this namespace will not be send to AppSignal.

Read more about namespaces.

log

Description

Select which logger the AppSignal integration will use. Accepted values are file and stdout. See also the log_path configuration.

• file (default)
  • Write all AppSignal logs to the file system.
• stdout (default on Heroku)
  • Print AppSignal logs in the parent process' STDOUT instead of to a file. Useful with hosting solutions such as container systems and Heroku.

log_level

Description

Set the severity level of AppSignal's internal logger. If it is configured to "info" it will log all error, warning and info messages, but not log the debug messages.

Setting it to the levels "debug" or "trace" is usually only needed on request from support. Setting the level to "debug"/"trace" could have a slight impact on the disk usage and IO, especially on high-traffic sites. CPU overhead is minimal with the debug option enabled.

Accepted values:

• error
• warning
• info
• debug
• trace

log_path

Description

Override the location of the path (directory) where the appsignal.log file can be written to.

nginx_port

Description

Configure the port on which the NGINX metrics server is exposed.

When AppSignal receives NGINX metrics, it listens on a localhost-bound server, by default on port 27649. If you're running several AppSignal-instrumented applications in the same server with NGINX metrics enabled, use this option to configure each application to listen on a different port.

opentelemetry_port

Description

Set this option to configure the OpenTelemetry HTTP server port of the AppSignal agent process. Configure this port if another process is already running on the machine that is also using this port to avoid conflicts.

request_headers

Description

The request_headers config option contains a list of HTTP request headers which are read and stored by the AppSignal Python package.

This request_headers config option is an allowlist, which means that it will only take headers as specified by this config option. If this config option is unset it will use the AppSignal default.

Following list is the AppSignal package default.

# __appsignal__.py
appsignal = Appsignal(
  # ...Other config opts
  request_headers=[
    "accept",
    "accept-charset",
    "accept-encoding",
    "accept-language",
    "cache-control",
    "connection",
    "content-length",
    "range"
  ]
)

To customize this list, add the config option with all headers you want reported. All header names should be written in lowercase characters. If the header name is X-Custom-Header, add it to this list as x-custom-header.

To configure AppSignal to not store any HTTP request headers on AppSignal traces, configure the option with an empty list.

# __appsignal__.py
appsignal = Appsignal(
  # ...Other config opts
  request_headers=[]
)

response_headers

Description

The response_headers config option contains a list of HTTP response headers which are read and stored by the AppSignal for Python package.

This response_headers config option is an allowlist, which means that it will only take headers as specified by this config option. If this config option is unset, it will not report any response headers.

To customize this list, add the config option with all headers you want reported. All header names should be written in lowercase characters. If the header name is X-Custom-Header, add it to this list as x-custom-header.

revision

Description

Set the app revision to report the currently running version of your app. AppSignal will create a deploy marker when this value changes, and tag all incoming data with the current revision.

When your application is deployed using Kamal, or when it is deployed to Render, or when it is deployed to Heroku and the Heroku Labs: Dyno Metadata feature is enabled, the AppSignal integration will automatically detect the Git commit of the current deployment and use it as the revision.

You can overwrite the automatically detected revisions in Heroku, Render or Kamal by manually setting the revision config option to a custom value.

Read more about deploy markers in the deploy markers topic.

running_in_container

Description

AppSignal expects to be running on the same machine between different deploys. Set this key to true if the application is running in a container, such as with Docker.

Newer versions of the AppSignal integration automatically detect its container environment, so no manual configuration is necessary. If you're having trouble with the automatic detection, please contact support.

This option is set to true automatically on Heroku.

send_environment_metadata

Description

Send environment metadata about the app.

For more information please read about environment metadata.

send_function_parameters

Description

Whether to send function parameters to AppSignal.

send_params

Description

Whether to skip sending request parameters to AppSignal.

For more information please read about send_params in filtering request parameters.

send_request_payload

Description

Whether to send request payload to AppSignal.

send_request_query_parameters

Description

Whether to send request query parameters to AppSignal.

send_session_data

Description

Set this option to false to not send any session data with exception traces and performance issue samples.

For more information please read about request session data filtering.

service_name

Description

The service name helps AppSignal recognize different services for traces and groups the traces automatically in a namespace based on the service name.

Choose a name that fits your application, like "Web server", "Background worker", "Email service", "Authentication service", etc.

statsd_port

Description

Set this option to configure the StatsD HTTP server port of the AppSignal agent process. Configure this port if another process is already running on the machine that is also using this port to avoid conflicts.

working_directory_path

Description

Override the location where AppSignal for Python can store temporary files. Use this option if the default location is not suitable. See our how AppSignal operates page for more information about the purpose of this working directory.

If you are running multiple applications using AppSignal on the same server, use this configuration option to select different working directories for every AppSignal instance, otherwise the two instances could conflict with one another. For more information on this scenario see our running multiple applications on one host documentation.

# __appsignal__.py
appsignal = Appsignal(
  # Other config
  working_directory_path="/tmp/project_1"
)