Configuration reference
This page lists the configuration properties of the Tock components.
Setting a property
Each property is read from a JVM system property or, failing that, from an environment variable of the same name. With Docker Compose for instance:
services:
admin_web:
image: tock/bot_admin:$TAG
environment:
- tock_env=prod
- tock_mongo_url=mongodb://mongo:27017/?replicaSet=tock
or on the command line: java -Dtock_env=prod ....
The value formats are:
- Durations: the unit is given by the suffix of the name (
_ms,_in_s,_in_minutes,_hours,_days...) - Lists: values separated by commas, e.g.
tock_users=alice@tock.ai,bob@tock.ai - Maps: entries separated by
|, keys and values separated by=, e.g.id1=namespace1|id2=namespace2
A property is only read by the components that embed the corresponding module: for instance, the connector
properties are read by the bot (bot_api service in Bot API mode, or your bot in Integrated bot mode).
This list is built from the source code with the
etc/list-doc-properties.pyscript, which lists the properties read by the Kotlin modules and the ones missing from this page (--checkoption).
Common properties
These properties are read by all the Tock components (bot_admin, nlp_api, build_worker, bot_api,
kotlin_compiler, duckling_bot) and by the bots developed in Integrated bot mode.
Environment
| Property | Default | Description |
|---|---|---|
tock_env |
dev |
Environment. Any value other than dev enables the production behaviors (HTTPS cookies, CORS, mandatory encryption passphrase...) |
tock_default_locale |
en |
Default locale of the platform |
tock_default_zone |
UTC |
Default time zone |
tock_default_namespace |
app |
Default namespace |
tock_namespace_open_access |
false |
Gives every user access to all the namespaces. Disabled by default for security reasons |
tock_vertx_worker_pool_size |
Vert.x default | Size of the Vert.x worker thread pool |
tock_cache_in_memory_maximum_size |
10000 |
Maximum number of entries of the in-memory caches |
tock_cache_in_memory_expiration_in_ms |
3600000 |
Expiration of the in-memory cache entries, after the last access |
MongoDB
See Installation.
| Property | Default | Description |
|---|---|---|
tock_mongo_url |
mongodb://localhost:27017,localhost:27018,localhost:27019/?replicaSet=tock&retryWrites=true |
MongoDB connection string. A replica set is required |
tock_async_mongo_url |
value of tock_mongo_url |
Connection string used by the asynchronous driver |
tock_bot_mongo_db |
tock_bot |
Name of the database of the bots (dialogs, users, stories, answers...) |
tock_front_mongo_db |
tock_front |
Name of the database of the NLP applications (intents, entities, sentences) |
tock_model_mongo_db |
tock_model |
Name of the database of the NLP models |
tock_cache_mongo_db |
tock_cache |
Name of the database of the shared cache |
tock_document_db_on |
false |
Enables the compatibility mode with Amazon DocumentDB |
tock_database_mongodb_secret_manager_provider |
Secret manager holding the MongoDB credentials: AWS_SECRETS_MANAGER or GCP_SECRET_MANAGER |
|
tock_database_mongodb_credentials_secret_name |
database_mongodb_credentials |
Name of the secret holding the MongoDB credentials |
Tock Studio authentication
See Tock Studio authentication. Without any of the *_enabled properties below, the users are
defined by the tock_users, tock_passwords, tock_organizations and tock_roles properties.
| Property | Default | Description |
|---|---|---|
tock_users |
value of tock_user |
Logins of the users |
tock_passwords |
value of tock_password |
Passwords of the users, in the same order |
tock_organizations |
value of tock_default_namespace |
Namespaces of the users, in the same order |
tock_roles |
all roles | Roles of the users, in the same order. The roles of a user are separated by \| |
tock_user |
admin@app.com |
Login of the default user, when tock_users is not set |
tock_password |
password |
Password of the default user, when tock_passwords is not set |
tock_oauth2_enabled |
false |
Enables the generic OAuth2 authentication |
tock_oauth2_client_id |
OAuth2 client identifier | |
tock_oauth2_secret_key |
OAuth2 client secret | |
tock_oauth2_site_url |
URL of the OAuth2 provider | |
tock_oauth2_access_token_path |
Relative path of the token endpoint | |
tock_oauth2_authorize_path |
Relative path of the authorization endpoint | |
tock_oauth2_userinfo_path |
Relative path of the user info endpoint | |
tock_oauth2_supported_grant_types |
Supported grant types | |
tock_oauth2_user_role_attribute |
custom:roles |
Attribute of the user info holding the roles |
tock_oauth2_proxy_host |
Host of the HTTP proxy used to reach the OAuth2 or GitHub provider | |
tock_oauth2_proxy_port |
0 |
Port of this proxy |
tock_oauth2_proxy_username |
User of this proxy | |
tock_oauth2_proxy_password |
Password of this proxy | |
tock_custom_roles_mapping |
Mapping between the roles of the OAuth2 provider and the Tock roles | |
tock_custom_namespace_mapping |
OAuth2: mapping between user profiles and Tock namespaces (e.g. id1=namespace1\|id2=namespace2). Keycloak: attribute holding the namespace (default tock_namespace) |
|
tock_bot_admin_rest_default_base_url |
http://localhost:8080 |
URL of Tock Studio, to redirect the user after the OAuth2 or Keycloak authentication |
tock_keycloak_enabled |
false |
Enables the Keycloak authentication |
tock_keycloak_client_id |
tock |
Keycloak client identifier |
tock_keycloak_secret_key |
Keycloak client secret | |
tock_keycloak_site_url |
https://keycloak/realms/myrealm |
URL of the Keycloak realm |
tock_keycloak_access_token_path |
/protocol/openid-connect/token |
Relative path of the token endpoint |
tock_keycloak_authorize_path |
/protocol/openid-connect/auth |
Relative path of the authorization endpoint |
tock_keycloak_userinfo_path |
/protocol/openid-connect/userinfo |
Relative path of the user info endpoint |
tock_keycloak_grant_types |
Supported grant types | |
tock_keycloak_user_role_attribute |
tock_roles |
Attribute holding the roles |
tock_keycloak_proxy_host |
Host of the HTTP proxy used to reach Keycloak | |
tock_keycloak_proxy_port |
0 |
Port of this proxy |
tock_github_oauth_enabled |
false |
Enables the GitHub authentication |
tock_github_oauth_client_id |
Client identifier of the GitHub OAuth application | |
tock_github_oauth_secret_key |
Client secret of the GitHub OAuth application | |
tock_github_api_request_timeout_ms |
5000 |
Timeout of the GitHub API calls |
tock_cas_auth_enabled |
false |
Enables the CAS authentication (requires a CAS authentication module) |
tock_cas_join_same_namespace_per_user |
true |
Puts the users of the same organization in the same namespace |
tock_cas_auth_proxy_host |
Host of the HTTP proxy used to reach the CAS server | |
tock_cas_auth_proxy_port |
Port of this proxy | |
aws_region |
eu-west-1 |
AWS region of the Application Load Balancer, for the authentication with its JWT |
tock_aws_public_key_request_timeout_ms |
30000 |
Timeout of the calls retrieving the AWS ALB public keys |
Web server
The properties prefixed by <verticle>_ apply to one HTTP server (verticle) of Tock. The prefix is the name of
the verticle class in lower case: botadminverticle (bot_admin), nlpverticle (nlp_api),
botverticle (bot and bot_api), kotlincompilerverticle (kotlin_compiler), ducklingverticle (duckling_bot),
healthcheckverticle (build_worker). For instance, botverticle_port sets the HTTP port of the bot.
| Property | Default | Description |
|---|---|---|
<verticle>_port |
8080 |
HTTP port |
<verticle>_body_limit |
1000000 |
Maximum size of a request body, in bytes |
<verticle>_tock_vertx_compression_supported |
true |
Enables HTTP compression |
<verticle>_tock_vertx_healthcheck_path |
<root path>/healthcheck |
Path of the healthcheck (see Supervision) |
<verticle>_tock_vertx_readinesscheck_path |
/health/readiness |
Path of the readiness probe |
<verticle>_tock_vertx_livenesscheck_path |
/health/liveness |
Path of the liveness probe |
tock_detailed_healthcheck_enabled |
false |
Returns the status of each dependency in the healthcheck (see Supervision) |
tock_https_env |
true |
Secure (HTTPS only) session cookies. Ignored in the dev environment |
tock_vertx_session_expiration_timeout |
21600000 |
Expiration of the Tock Studio sessions, in milliseconds (6 hours) |
tock_web_use_default_cors_handler |
true in dev, else false |
Adds a CORS handler to the HTTP servers |
tock_web_use_default_cors_handler_url |
http://localhost:4200 (* for the bot) |
Origins allowed by this CORS handler, separated by \| (*: all) |
tock_web_use_default_cors_handler_with_credentials |
true (false for the bot) |
Allows the credentials in the CORS requests |
tock_web_cookie_auth |
Default user identification of the web connectors: encrypted (encrypted cookie), basic or true (plain cookie). Otherwise, the identifier sent by the client is used |
|
tock_web_cookie_auth_max_age |
-1 |
Maximum age of the authentication cookie, in seconds (-1: session cookie) |
tock_web_cookie_auth_path |
Path of the authentication cookie | |
tock_web_sse_keepalive_delay |
10 |
Delay between two keep-alive messages of the SSE connections, in seconds |
tock_file_upload_directory |
file-uploads |
Directory of the uploaded files |
Encryption and secrets
See Security.
| Property | Default | Description |
|---|---|---|
tock_encrypt_pass |
Passphrase used to encrypt the sensitive data. Required outside the dev environment as soon as encryption is used | |
tock_gen_ai_secret_manager_provider |
Secret manager holding the Gen AI secrets (API keys): AWS_SECRETS_MANAGER or GCP_SECRET_MANAGER |
|
tock_gen_ai_secret_prefix |
LOCAL/TOCK |
Prefix of the names of the Gen AI secrets |
Logs and outgoing HTTP calls
See Supervision.
| Property | Default | Description |
|---|---|---|
tock_logback_enabled |
true |
Enables the default Logback configuration of Tock |
tock_default_log_level |
DEBUG in dev, else INFO |
Default log level |
tock_logback_file_appender |
false |
Also writes the logs to log/logFile.log |
tock_retrofit_log_level |
BODY in dev, else NONE |
Log level of the outgoing HTTP calls: NONE, BASIC, HEADERS or BODY |
tock_proxy_user |
User of the HTTP proxy used for the outgoing calls | |
tock_proxy_password |
Password of this proxy | |
tock_circuit_breaker |
false |
Enables a circuit breaker on the outgoing calls that support it (Duckling) |
Bot
These properties are read by the bots: the bot_api service in Bot API mode, or your bot application
in Integrated bot mode.
Bot engine
| Property | Default | Description |
|---|---|---|
tock_configuration_bot_default_base_url |
http://<local IP>:<botverticle_port> |
Base URL of the bot, proposed by default in the connector configurations of Tock Studio |
tock_bot_protected_paths |
/admin |
Paths of the bot protected by the authentication |
tock_bot_protected_path |
/admin |
Deprecated: use tock_bot_protected_paths |
tock_restricted_configuration_id |
Only loads the given bot configurations (development) | |
tock_bot_default_breath_ms |
1000 |
Default delay between two messages of the bot |
tock_bot_min_breath_ms |
50 |
Minimum delay between two messages, when a message takes longer than the default delay to prepare |
tock_technical_error |
Technical error :( sorry! |
Message sent to the user on technical error |
tock_gen_ai_orchestrator_technical_error |
Technical error :( sorry! |
Message sent to the user when the Gen AI orchestrator fails |
tock_ask_again_round |
1 |
Default number of times the bot asks a question again |
tock_bot_send_choice_activate |
true |
A click on a button reactivates a disabled bot |
tock_bot_disabled_duration_in_minutes |
7200 |
Duration of the deactivation of the bot for a user (5 days) |
tock_bot_refresh_profil_duration_in_minutes |
7200 |
Delay before refreshing the user profile from the channel (5 days) |
tock_bot_wait_nlp_availability_in_ms |
5000 |
Time to wait for the NLP service at startup |
tock_cleanup_delay_seconds |
60 |
Maximum delay to send the queued messages of a coroutine story |
tock_bot_audio_nlp_enabled |
true |
Sends the audio messages to the NLP (speech to text) |
tock_bot_audio_nlp_max_size |
1048576 |
Maximum size of the audio messages sent to the NLP, in bytes |
tock_bot_serve_files |
true |
Serves the files uploaded in Tock Studio (answer images...) |
tock_bot_serve_files_path |
/f/ |
Path of these files |
tock_nlp_proxy_on_bot |
false |
Exposes the NLP API through the bot |
tock_nlp_proxy_on_bot_path |
/_proxy_nlp |
Path of this NLP proxy |
tock_websocket_enabled |
false |
Enables the WebSocket endpoint of the Bot API mode |
tock_timeline_persistence_synchronous_mode |
true |
Deprecated: use coroutine stories |
tock_timeline_persistence_asynchronous_mode |
false |
Deprecated: use coroutine stories |
tock_orchestration |
false |
Enables the orchestration between bots |
tock_orchestration_lock_ttl_in_min |
60 |
Lifetime of the orchestration locks |
tock_gen_ai_orchestrator_server_url |
http://localhost:8000 |
URL of the Gen AI orchestrator |
tock_gen_ai_orchestrator_client_request_timeout_ms |
55000 |
Timeout of the calls to the Gen AI orchestrator |
tock_gen_ai_orchestrator_vector_store_provider |
PGVector |
Type of the default vector store (see Vector DB settings) |
tock_nlp_client_request_timeout_ms |
20000 |
Timeout of the calls to the NLP API |
Test keywords and test user
Keywords sent as a user message to act on the conversation, mostly used for the tests.
| Property | Default | Description |
|---|---|---|
tock_bot_delete_keyword |
_delete_user_ |
Deletes the data of the user |
tock_bot_enable_keyword |
_enable_user_ |
Enables the bot for the user |
tock_bot_disable_keyword |
_disable_user_ |
Disables the bot for the user |
tock_bot_test_context_keyword |
_test_ |
Starts a test context |
tock_bot_end_test_context_keyword |
_end_test_ |
Ends the test context |
tock_bot_test_first_name |
Joe |
First name of the test user |
tock_bot_test_last_name |
Hisaishi |
Last name of the test user |
Dialog storage
See Installation.
| Property | Default | Description |
|---|---|---|
tock_bot_timeline_index_ttl_days |
365 |
Retention of the user timelines (profiles and preferences) |
tock_bot_dialog_index_ttl_days |
7 |
Retention of the dialogs |
tock_bot_flow_stats_index_ttl_days |
365 |
Retention of the conversation flow statistics |
tock_bot_alternative_index_ttl_hours |
1 |
Retention of the alternatives chosen for the multiple-choice answers |
tock_bot_dialog_max_validity_in_seconds |
86400 |
Maximum age of a dialog to be continued. After this delay, a new dialog starts |
tock_bot_max_actions_by_dialog |
1000 |
Maximum number of actions kept in a dialog |
tock_bot_dialog_flow_stat |
true |
Records the conversation flow statistics |
tock_dialog_flow_crawl_stats |
true |
Computes the conversation flow statistics periodically |
tock_bot_add_namespace_to_timeline_id |
false |
Prefixes the user identifiers with the namespace, to separate the users of several namespaces |
tock_bot_encrypted_flags |
Dialog flags whose value is encrypted | |
tock_bot_lock_timeout_in_ms |
5000 |
Lifetime of the lock of a user, while a message is processed |
tock_bot_locked_attempts_wait_in_ms |
500 |
Delay between two attempts to acquire the lock |
tock_bot_lock_retry_jitter_in_ms |
20 |
Random delay added to this delay |
tock_bot_max_locked_attempts |
10 |
Number of attempts before giving up |
tock_bot_warn_after_locked_attempts |
5 |
Number of attempts before logging a warning |
mongo_user_ttl_hours |
6 |
Retention of the user locks |
Bot API mode
See Bot API.
| Property | Default | Description |
|---|---|---|
tock_bot_api_timeout_in_ms |
60000 |
Timeout of the calls to the webhook of the bot |
tock_bot_api_connection_timeout_in_ms |
3000 |
Connection timeout of these calls |
tock_bot_api_webhook_check_reachability |
true |
Checks that the webhook is reachable before calling it |
tock_bot_api_webhook_reachability_in_ms |
10000 |
Timeout of this check |
tock_api_timout_in_s |
10 |
Timeout of the answer of a bot connected with WebSocket |
tock_api_old_webhook_behaviour |
false |
Disables the streaming (SSE) of the webhook answers |
tock_bot_api_actions_history_to_client_bus |
false |
Sends the history of the dialog actions to the bot with each request |
tock_websocket_host |
localhost |
Host of the bot_api service, for the WebSocket client |
tock_websocket_port |
8080 |
Port of the bot_api service, for the WebSocket client |
tock_websocket_ssl |
false |
Uses TLS for the WebSocket connection |
Connectors
Web
See Web.
| Property | Default | Description |
|---|---|---|
tock_web_sse |
false |
Enables the SSE (Server-Sent Events) endpoint |
tock_web_direct_sse |
false |
Sends the answers directly on the SSE connection, without storing them |
tock_web_cors_pattern |
.* |
Origins allowed to call the connector |
tock_web_enable_markdown |
false |
Converts the Markdown of the answers to HTML |
allow_markdown |
Former name of tock_web_enable_markdown |
|
tock_web_connector_persist_profile |
false |
Stores the user profile sent by the client |
tock_web_connector_extra_headers |
HTTP headers of the request passed to the bot | |
tock_web_connector_use_extra_header_as_metadata_request |
false |
Passes these headers as metadata of the request |
tock_web_connector_merge_stream_response |
true |
Merges the chunks of a streamed answer into one message |
tock_web_connector_bridge_enabled |
false |
Allows a request to target another web connector with its connectorId field |
tock_web_sse_message_queue_ttl_days |
-1 |
Retention of the SSE messages waiting to be delivered (-1: no expiration) |
tock_web_sse_message_queue_max_count |
50000 |
Maximum number of SSE messages waiting to be delivered |
tock_web_sse_message_queue_max_size_kb |
100000 |
Maximum size of these messages, in kilobytes |
OpenAI API
See OpenAI API. This connector also reads tock_web_sse and tock_web_cors_pattern.
| Property | Default | Description |
|---|---|---|
tock_openai_connector_persist_profile |
false |
Stores the user profile sent by the client |
tock_openai_support_unstreamed |
false |
Accepts the requests without streaming |
See WhatsApp.
| Property | Default | Description |
|---|---|---|
tock_whatsappcloud_request_timeout_ms |
30000 |
Timeout of the calls to the WhatsApp API |
tock_whatsappcloud_request_gzip |
false |
Compresses the requests |
tock_whatsapp_cloud_restricted_phone_numbers |
Only answers these phone numbers (tests) | |
tock_whatsapp_cloud_restricted_user_ids |
Only answers these Business-Scoped User IDs (BSUID), in addition to the phone numbers above (tests) | |
tock_whatsapp_sync_templates |
false |
Synchronizes the message templates with WhatsApp |
tock_whatsapp_reupload_images |
true |
Uploads the images to WhatsApp instead of sending their URL |
tock_whatsapp_error_on_invalid_messages |
false |
Fails instead of logging a warning on a message that does not respect the WhatsApp limits |
tock_whatsapp_payload |
whatsapp_payload |
Collection storing the button payloads |
tock_whatsapp_payload_ttl_days |
10 |
Retention of these payloads |
Messenger
See Messenger.
| Property | Default | Description |
|---|---|---|
tock_messenger_request_timeout_ms |
30000 |
Timeout of the calls to the Messenger API |
tock_messenger_request_gzip |
true |
Compresses the requests |
messenger_retries_on_error_limit |
1 |
Number of retries of a failed call |
messenger_retries_on_error_wait_in_ms |
5000 |
Delay between two retries |
tock_messenger_extended_profile_fields |
Additional fields of the user profile to retrieve | |
tock_bot_messenger_reuse_attachment |
true |
Reuses the attachments already uploaded to Messenger |
tock_messenger_webhook_check_subscription |
false |
Periodically checks the webhook subscription of the pages |
tock_messenger_webhook_check_period |
600 |
Period of this check, in seconds |
tock_messenger_webhook_url |
URL of the current subscription | Webhook URL used to renew the subscription |
Other connectors
| Property | Default | Description |
|---|---|---|
tock_microsoft_request_timeout |
5000 |
Teams: timeout of the calls retrieving the Microsoft keys, in milliseconds |
tock_teams_request_timeout_ms |
30000 |
Teams: timeout of the calls to the Bot Framework API (formerly tock_whatsapp_request_timeout_ms, still read if the new property is not set) |
tock_whatsapp_request_timeout_ms |
30000 |
WhatsApp (deprecated On-Premise API connector): timeout of the calls to the WhatsApp API |
tock_mattermost_request_timeout_ms |
30000 |
Mattermost: timeout of the calls to the Mattermost API |
tock_slack_old_api_style |
false |
Slack: uses the former format of the Slack API |
tock_api_google_chat_connector_test_send_intro_message |
false |
Google Chat: sends an introduction message in test mode |
tock_rest_connector_disabled |
false |
Disables the REST connector used by the tests of Tock Studio |
tock_rest_connector_check_nlp_stats |
false |
REST connector: returns the NLP statistics with the answer |
tock_bot_rest_client_request_timeout_ms |
100000 |
Timeout of the calls of the REST connector client |
tock_iadvize_secret_manager_provider |
ENV |
iAdvize: secret manager holding the credentials |
tock_iadvize_credentials_secret_name |
iadvize_credentials |
iAdvize: name of the secret holding the credentials |
Tock Studio
These properties are read by the bot_admin service.
| Property | Default | Description |
|---|---|---|
botadminverticle_base_href |
/ |
Base path of Tock Studio (see Installation) |
botadminverticle_content_path |
/maven/dist |
Directory of the Tock Studio web application |
botadminverticle_nlp_external_host |
localhost:8888 |
Host of the NLP API displayed in Tock Studio |
tock_bot_api |
false |
Displays the Bot API features in Tock Studio |
tock_bot_global_message |
Message displayed in a banner of Tock Studio | |
tock_csv_delimiter |
; |
Delimiter of the CSV exports |
tock_bot_compiler_disabled |
false |
Disables the script compiler |
tock_bot_compiler_service_url |
http://localhost:8887 |
URL of the kotlin_compiler service |
tock_bot_compiler_timeout_in_ms |
60000 |
Timeout of the calls to this service |
tock_kotlin_compiler_classpath |
Additional classpath of the kotlin_compiler service |
|
tock_bot_test_xray_url |
URL of Xray: enables the Xray features of the test plans |
Answers (i18n)
See Building a multilingual bot.
| Property | Default | Description |
|---|---|---|
tock_i18n_enabled |
false |
Enables the multilingual answers |
tock_load_all_labels_of_default_namespace |
true |
Loads all the labels of the default namespace at startup |
tock_i18n_stat_write_enabled |
true |
Records the usage statistics of the labels |
tock_i18n_stat_refresh_in_ms |
10000 |
Period of the recording of these statistics |
tock_i18n_reset_value_on_default_change |
false |
Resets the translations of a label when its default value changes |
tock_translator_deepl_api_key |
DeepL API key (tock-deepl-translate module) |
|
tock_translator_deepl_api_url |
https://api.deepl.com/v2/translate |
URL of the DeepL API |
tock_translator_deepl_target_languages |
all | Target languages supported by DeepL |
tock_translator_deepl_glossary_map_ids |
DeepL glossaries, by language |
NLP
These properties are read by the nlp_api, bot_admin and build_worker services.
| Property | Default | Description |
|---|---|---|
tock_nlp_root |
/rest/nlp |
Root path of the NLP API |
nlpverticle_tock_nlp_protect_path |
false |
Protects the NLP API with the authentication |
nlpverticle_tock_nlp_check_healthcheck |
false |
Checks the entity providers in the healthcheck |
tock_nlp_model_refresh |
true |
Reloads the models when they are rebuilt |
tock_nlp_model_fill_cache |
false |
Loads all the models at startup |
tock_parser_validate_sentence_test |
false |
In test mode, returns the validated qualification of a known sentence |
nlp_duckling_url |
http://localhost:8889 |
URL of the Duckling service |
tock_duckling_enabled |
true |
Enables the Duckling entities (dates, numbers...) |
tock_duckling_request_timeout_ms |
4000 |
Timeout of the calls to Duckling |
tock_nlp_entity_type_url |
http://localhost:5000/app/v1/ |
URL of the REST entity provider (see NLP and entity models) |
tock_bot_rest_timeout_in_ms |
10000 |
Timeout of the calls to this provider |
rasa_model_path |
models/ |
Directory of the Rasa models |
Data retention
See Installation.
| Property | Default | Description |
|---|---|---|
tock_nlp_log_index_ttl_days |
7 |
Retention of the NLP request logs |
tock_nlp_log_stats_index_ttl_days |
365 |
Retention of the NLP statistics |
tock_user_log_index_ttl_days |
365 |
Retention of the logs of the Tock Studio user actions |
tock_nlp_classified_sentences_index_ttl_days |
-1 |
Retention of the sentences not validated (-1: no expiration) |
tock_nlp_classified_sentences_index_ttl_intent_names |
Limits this retention to the sentences of these intents |
Model building
| Property | Default | Description |
|---|---|---|
tock_build_worker_mode |
VERTICLE |
Execution mode of the build_worker: VERTICLE (service), COMMAND_LINE, ON_DEMAND or DEV |
tock_build_worker_verticle_enabled |
true |
Deprecated: use tock_build_worker_mode |
tock_complete_model_enabled |
true |
Periodically rebuilds the complete models |
tock_test_model_enabled |
true |
Runs the model tests (see Model quality) |
tock_test_model_timeframe |
0,5 |
Hours (start, end) during which the model tests run |
tock_cleanup_model_enabled |
true |
Deletes the models that are no longer used |
tock_build_worker_on_demand_type |
AWS_BATCH |
ON_DEMAND mode: type of the jobs |
tock_build_worker_on_demand_delay_in_minutes_between_job_rebuild_diff |
60 |
ON_DEMAND mode: delay between two model updates |
tock_build_worker_on_demand_timeframe_rebuild_diff |
0,24 |
ON_DEMAND mode: hours during which the models are updated |
tock_build_worker_on_demand_delay_in_minutes_between_job_test |
1440 |
ON_DEMAND mode: delay between two model tests |
tock_build_worker_on_demand_timeframe_test |
0,5 |
ON_DEMAND mode: hours during which the models are tested |
tock_build_worker_on_demand_delay_in_minutes_between_job_cleanup |
720 |
ON_DEMAND mode: delay between two cleanups |
tock_build_worker_on_demand_timeframe_cleanup |
0,24 |
ON_DEMAND mode: hours during which the cleanups run |
tock_worker_aws_batch_job_definition_name |
tock-worker-job-definition |
AWS Batch: job definition |
tock_worker_aws_batch_job_queue_name |
tock-worker-job-queue |
AWS Batch: job queue |
tock_worker_aws_batch_job_name |
tock-worker-job |
AWS Batch: prefix of the job names |
tock_worker_aws_batch_attempt_duration_seconds |
7200 |
AWS Batch: maximum duration of a job |
tock_worker_aws_batch_vcpus |
4 |
AWS Batch: vCPUs of a job |
tock_worker_aws_batch_memory |
12288 |
AWS Batch: memory of a job, in MB |
Amazon SageMaker models
| Property | Default | Description |
|---|---|---|
tock_sagemaker_aws_region_name |
eu-west-3 |
AWS region of the SageMaker endpoints |
tock_sagemaker_aws_profile_name |
default |
AWS profile |
tock_sagemaker_aws_intent_endpoint_name |
default |
Endpoint of the intent model |
tock_sagemaker_aws_entities_endpoint_name |
default |
Endpoint of the entity model |
tock_sagemaker_aws_content_type |
application/json |
Content type of the requests |
tock_sagemaker_aws_model_file_name |
default |
Name of the model file |
Gen AI orchestrator
The Python Gen AI orchestrator reads environment variables prefixed by tock_gen_ai_orchestrator_.
The vector store variables are described on the Vector DB settings page.
| Variable | Default | Description |
|---|---|---|
tock_gen_ai_orchestrator_application_environment |
DEV |
DEV or PROD |
tock_gen_ai_orchestrator_application_logging_config_ini |
provided file | Logging configuration file |
tock_gen_ai_orchestrator_llm_provider_timeout |
30 |
Timeout of the LLM calls, in seconds |
tock_gen_ai_orchestrator_llm_provider_max_retries |
0 |
Number of retries of a failed LLM call |
tock_gen_ai_orchestrator_llm_rate_limits |
true |
Enables the rate limit of the LLM calls |
tock_gen_ai_orchestrator_em_provider_timeout |
4 |
Timeout of the embedding calls, in seconds |
tock_gen_ai_orchestrator_compressor_provider_timeout |
7 |
Timeout of the document compressor calls, in seconds |
tock_gen_ai_orchestrator_observability_provider_timeout |
3 |
Timeout of the observability calls, in seconds |
tock_gen_ai_orchestrator_observability_provider_max_retries |
0 |
Number of retries of a failed observability call |
tock_gen_ai_orchestrator_db_pool_size |
10 |
Size of the PGVector connection pool |
tock_gen_ai_orchestrator_db_max_overflow |
5 |
Additional connections allowed beyond this size |
tock_gen_ai_orchestrator_db_pool_timeout |
30 |
Timeout to get a connection, in seconds |
tock_gen_ai_orchestrator_db_pool_recycle |
3600 |
Maximum lifetime of a connection, in seconds |
tock_gen_ai_orchestrator_vector_store_test_query |
Any definition |
Query used to test the connection to the vector store |
tock_gen_ai_orchestrator_vector_store_test_max_docs_retrieved |
4 |
Number of documents retrieved by this test |
tock_gen_ai_orchestrator_aws_bedrock_credentials_profile_name |
AWS profile used to call AWS Bedrock | |
tock_gen_ai_orchestrator_aws_bedrock_credentials_allow_default_profile |
false |
Without profile, uses the default AWS credential chain (IAM role, IRSA, environment variables) instead of failing |
tock_gcp_project_id |
GCP project of the secrets, with the GCP secret manager |
Other modules
Xray tests
Properties of the tock-xray-plugin module, which runs the Xray test plans of a bot.
| Property | Default | Description |
|---|---|---|
tock_bot_test_url |
URL of Tock Studio | |
tock_bot_test_login |
Tock Studio login | |
tock_bot_test_password |
Tock Studio password | |
tock_bot_test_timeout_in_ms |
3600000 |
Timeout of the calls to Tock Studio |
tock_bot_test_botId |
Identifier of the tested bot | |
tock_bot_test_configuration_ids |
Identifiers of the tested configurations | |
tock_bot_test_configuration_names |
Names of the tested configurations | |
tock_bot_test_locale |
value of tock_default_locale |
Locale of the tests |
tock_bot_test_start_sentence |
Sentence sent at the beginning of each test | |
tock_bot_test_xray_login |
Xray login | |
tock_bot_test_xray_password |
Xray password | |
tock_bot_test_xray_timeout_in_ms |
60000 |
Timeout of the calls to Xray |
tock_bot_test_xray_test_plan_keys |
Keys of the test plans to run | |
tock_bot_test_xray_test_keys |
Keys of the tests to run | |
tock_bot_test_xray_test_plan_env |
Environment reported in the test executions | |
tock_bot_test_xray_test_plan_bot_url |
Bot URL reported in the test executions | |
tock_bot_test_jira_project |
Jira project of the tests | |
tock_bot_test_jira_xray_test_type_field |
Jira field of the test type | |
tock_bot_test_jira_xray_manual_test_field |
Jira field of the manual test steps | |
tock_bot_test_jira_xray_automation_type_field |
Jira field of the automation type | |
tock_bot_test_jira_linked_field |
Jira field linking the tests | |
tock_bot_test_jira_registered_bot_url |
${botUrl} |
Bot URL recorded in Jira |
tock_bot_test_connector_jira_map |
Jira components by connector | |
tock_bot_test_precondition_key_text_chat |
Xray precondition of the text tests | |
tock_bot_test_precondition_key_voice_assistant |
Xray precondition of the voice tests | |
tock_bot_test_precondition_key_text_and_voice_assistant |
Xray precondition of the text and voice tests | |
tock_bot_xray_creation_keyword |
_xray_ |
Keyword creating an Xray test from the current dialog |
tock_bot_xray_update_keyword |
_xray_update_ |
Keyword updating an Xray test from the current dialog |
Dialogflow
| Property | Default | Description |
|---|---|---|
dialogflow_project_id |
Google Cloud project of the Dialogflow agent to import |
Chat with Tock