config.md

Configuring Frontik

Here is the list of all options currently supported by Frontik. All of them can be set in the configuration file (see frontik_dev.cfg.ex for example), but you can pass them to /usr/bin/frontik script through command line as well. The exception is config option — it makes sense to pass it only through command line.

These options can be set for each Frontik instance (see options.py).

Option name	Type	Default value	Description
host	str	'0.0.0.0'	Host value for incoming connections
port	int	8000	Port to listen to
reuse_port	bool	True	Binds server socket with SO_REUSEPORT option
config	str	None	Path to config file
slow_callback_threshold_ms	int	None	Enables logging of long-running callbacks
app	str	None	Application name (is used for application identification and as default path to it's class)
app_class	str	None	Path to application class (by default FrontikApplication class is used (see Frontik application structure))
workers	int	1	Number of worker processes creates using fork. When default value is used, master itself become worker, without fork
init_workers_timeout_sec	int	60	Timeout for worker initialization
xheaders	bool	False	Controls Tornado HTTPServer xheaders option
tornado_settings	dict	None	tornado.web.Application settings
autoreload	bool	False	Restart Frontik after changes in application sources or config files
debug	bool	False	Enable debug mode
debug_login	str	None	Debug mode login for basic authentication (when debug=False)
debug_password	str	None	Debug mode password for basic authentication (when debug=False)
max_active_handlers	int	100	Limit the maximum number of simultaneous requests per worker
datacenter	str	None	Datacenter where current application is running
validate_request_id	bool	False	Enable check for validity of request_id (32 characters hex-string)
node_name	str	False	Name of node where frontik application starts
common_executor_pool_size	int	10	The maximum number of threads that can be used to execute the given calls
stop_timeout	int	3	Max time in seconds for gracefully application stopping

Logging options:

Option name	Type	Default value	Description
log_level	str	info	Python log level
logformat	str	see code	Log entry format for files and syslog
log_dir	str	None	Log directory location (set to None to disable logging to file)
log_json	bool	True	Enable JSON logging for files and syslog
log_text_format	str	see code	Log format for files and syslog when JSON logging is disabled
stderr_log	bool	False	Send log output to stderr (colorized if possible)
stderr_format	str	see code	Log entry format for stderr output
stderr_dateformat	str	see code	Log entry date format for stderr output
syslog	bool	False	Enables sending logs to syslog
syslog_host	str	127.0.0.1	Syslog host
syslog_port	int	None	Syslog port. If this value is None, unix socket is used, UDP otherwise
syslog_facility	str	'user'	Syslog facility
syslog_tag	str	''	Syslog tag
suppressed_loggers	list	[]	List of logger names to be excluded from debug output
sentry_dsn	str	None	Enable Sentry and set Sentry DSN for sending errors
statsd_host	str	None	Stats server host for metrics
statsd_port	int	None	Stats server port for metrics
statsd_default_periodic_send_interval_sec	int	60	Stats default periodic metrics sending interval
asyncio_task_threshold_sec	float	None	Threshold for logging long-running asyncio tasks
asyncio_task_critical_threshold_sec	float	None	Threshold for send to Sentry long-running asyncio tasks

HTTP client options:

Option name	Type	Default value	Description
max_http_clients	int	100	Curl max clients option
max_http_clients_connects	int	None	Curl max connects option
http_client_default_connect_timeout_sec	float	0.2	Default connect timeout
http_client_default_request_timeout_sec	float	2.0	Default request timeout
http_client_default_max_tries	int	2	Maximum number of retries per request + 1
http_client_default_max_timeout_tries	int	1	Maximum number of retries due to timeout per request + 1
http_client_default_retry_policy	dict	{599: False, 503: False}	Conditions when request retry is possible
http_proxy_host	str	None	HTTP proxy host for Curl HTTP client
http_proxy_port	int	3128	HTTP proxy port for Curl HTTP client
http_client_allow_cross_datacenter_requests	bool	False	Allow requests to different datacenter when no upstream in current datacenter is available
timeout_multiplier	float	1.0	Generic timeout multiplier for http requests (useful for testing)

Producers options:

Option name	Type	Default value	Description
xsl_root	str	None	Root directory for XSL files
xml_root	str	None	Root directory for XML files
xsl_cache_limit	int	None	Upper limit for XSL LRU files cache
xml_cache_limit	int	None	Upper limit for XML LRU files cache
xsl_cache_step	int	None	Increase in weight for XSL cache entry after each get
xml_cache_step	int	None	Increase in weight for XML cache entry after each get
xsl_executor_pool_size	int	1	Number of background threads for XSLT processing
jinja_template_root	str	None	Root directory for Jinja templates
jinja_template_cache_limit	int	50	Upper limit for Jinja templates cache
jinja_streaming_render_timeout_ms	int	50	Upper limit (in msecs) for one iteration of partial Jinja template rendering

Consul options:

Option name	Type	Default value	Description
consul_enabled	bool	True	Enable Consul features: registration, kv-store, upstreams
consul_host	str	127.0.0.1	Consul host
consul_port	int	None	Consul port
consul_service_address	str	None	Address of application for Consul registration
consul_check_host	str	None	Address for healthcheck application, if not provided, consul_service_address is used
consul_http_check_interval_sec	int	10	Interval for making healthcheck request
consul_http_check_timeout_sec	int	1	Timeout for health check request
consul_tags	list	[]	Additional parameters, which will will be sent with Consul registration
consul_weight_watch_seconds	int	600	Max waiting blocking query time for get host weight from Consul
consul_weight_total_timeout_sec	int	650	Timeout for waiting blocking query for getting host weight from Consul
consul_cache_initial_warmup_timeout_sec	int	2	Ttl of HTTP session to initialize cache
consul_cache_backoff_delay_seconds	int	10	Interval to wait until next attempt to fetch info from consul after exception
consul_consistency_mode	str	default	May set as default, consistent and stale. See Consul documentation
consul_weight_consistency_mode	str	default	Consistency mode for get weigth.
consul_deregister_critical_timeout	str	120h	Time, after that service will be de-registered automatically from Consul
upstreams	list	[]	List of service upstreams - services, where current app will send http requests
fail_start_on_empty_upstream	bool	True	If True app will not start if one or more upstreams don't have host addresses

Opentelemetry options:

Option name	Type	Default value	Description
opentelemetry_collector_url	str	127.0.0.1	OpenTelemetry collector url
opentelemetry_sampler_ratio	float	0.01	Probability (between 0 and 1) that a span will be sampled
opentelemetry_enabled	bool	False	Enable OpenTelemetry features

The only option that is mandatory for running Frontik is app option — the name of application package.

There are also certain options, that can be defined during application initialization, see Configuring Frontik application for more details.