NEXEED IAS | Operations Manual

Yes

database credentials (Can be used with and without prefix for global configuration. "with prefix" takes precedence)

Yes

database credentials (Can be used with and without prefix for global configuration. "with prefix" takes precedence)

Yes

JDBC connection url to the database

Yes

JDBC Driver class to connect to the database

Yes

host on which RabbitMQ is listening

Yes

host on which RabbitMQ is listening

Yes

port on which RabbitMQ is listening

Yes

user which should be used to get access to RabbitMQ

No

/

N0

Defines whether communication with RabbitMQ is TLS encrypted

true

No

Authentication method used when connecting to RabbitMQ. Can be one of 'EXTERNAL', 'AMQPLAIN', or 'PLAIN'

PLAIN

No

"Can be used to manually disable TLS for HTTP endpoint. If set to true, cert.pem and key.pem do not need to be mounted to container"

Yes

MACMA Url used for OAuth2

Yes

MMPD Url used for OAuth2

Yes

PORTAL base Url used for OAuth2

Yes

MACMA OAuth2 Client Scope

Yes

MDM OAuth2 Client Scope

Yes

Portal OAuth2 Client Scope

Yes

Tenant Zero required for Macma Migration of the Module

No

Optionally disable TLS for Java for specific use-cases such as development testing

false

No

Filename including location of the trusted certificate file (cacerts)

/etc/ssl/java/cacerts

No

Trust store password

changeit

No

Key password (cannot be different than trust store password?)

changeit

No

Certificate file for server TLS endpoint

server.crt

No

Key file for server TLS endpoint in /etc/ssl/server

server.key

No

Optional password for key file in /etc/ssl/server

secret

No

Optional additional trusted certificate in

/etc/ssl/trusted

No

Filename of keystore (relative to certificate volume)

/etc/ssl/java/java.pfx

No

Key store password

changeit

No

used for logging

No

used for logging

ai-core

No

used for logging

unknown

No

ID of the Microservice. Used for logging and health details.

NEXEED_GLOBAL_APPLICATION_INSTANCE_ID_example

No

"Maximum number of unique process types for which datasets are simultaneously aggregated. This will have linear dependency on the maximum required storage space on the database. More slots will result in faster collection and training of new models. Currently each tenant is allowed to collect data for this number of unique processes."

20

No

"Cron to trigger eviction of model execution data. Default means every 10 absolute seconds e.g. 10:00:10 10:00:20"

0/10 * * * * *

No

"Will be used to provide internet access for data bricks."

None

No

"The DATABRICKS_HOST_URL environment variable stores the web address of a Databricks workspace or cluster for programmatic access and management."

None

No

"Databricks token environment variable is a secure way to store and access authentication tokens for Databricks services."

None

No

General logging for ai-core.

WARN

No

General logging for the packages developed under Bosch.

WARN

No

Logging for the AMQP messages ai-core.

WARN

No

Logging for the Spring Framework messages in ai-core.

WARN

No

Logging for the Spring Security messages in ai-core.

INFO

No

Logging for the Spring Web messages in ai-core.

WARN

No

Logging for the Object Relational Mapper in ai-core.

WARN

No

Number of open DB connections in the pool

15

No

Number of days result log records for Anomaly Detection remains in DB

2

No

Number of days result log records for Sequence Detection remains in DB

2

Yes

host on which RabbitMQ is listening

Yes

host on which RabbitMQ is listening

Yes

port on which RabbitMQ is listening

Yes

user which should be used to get access to RabbitMQ

Yes

In multi-tenant environments, use a separate vhost for each tenant/environment

Yes

Defines whether communication with RabbitMQ is TLS encrypted

true

No

Authentication method used when connecting to RabbitMQ. Can be one of 'EXTERNAL', 'AMQPLAIN', or 'PLAIN'

PLAIN

No

IP Interface Web server listens to (Currently only used for health endpoint)

0.0.0.0

No

Port used for HTTP socket (Currently only used for health endpoint)

8000

No

"Can be used to manually disable TLS for HTTP endpoint. If set to true, cert.pem and key.pem do not need to be mounted to container"

false

No

Port used for HTTPS Socket (Currently only used for health endpoint)

8443

Yes

"Certificate used by Server during TLS connections. Files need to be mounted to '/etc/ssl/server/{{SERVER_CERT_FILE}}'"

cert.pem

Yes

"Keyfile used by server during TLS connections. Files need to be mounted to '/etc/ssl/server/{{SERVER_KEY_FILE}}'"

key.pem

No

used to toggle Elastic APM on / off

true

Yes

URL of Elastic instance for OpenTelemetry exporter

Yes

used for authentication with Elastic

Yes

used for authentication and exporting OpenTelemetry to Elastic using ELASTIC_APM_SECRET_TOKEN

No

Elastic APM Log-Level

INFO

No

Specifies which exporter is used for logs. Possible values are 'none', 'otlp'.

otlp

Yes

Necessary to reduce the load. Specifies the Sampler used to sample traces by the SDK.

"parentbased_always_on"

Yes

Necessary to reduce the load. Specifies arguments, if applicable, to the sampler defined in by OTEL_TRACES_SAMPLER. The specified value will only be used if OTEL_TRACES_SAMPLER is set.

none

Yes

"OPTION 1: TNS-connection string → (driver oracle+oracledb will be used) OPTION 2: jdbc-url → (python driver will be inferred from jdbc-schema ['oracle:thin:@//', 'sqlserver://', 'postgresql://']). jdbc-url is only accepted in this format: 'jdbc:sqlserver://mssql.db.svc.cluster.local:1433/ppm' OPTION 3: sqlalchemy compatible connection string (Can be used with and without prefix for global configuration. ""with prefix"" takes precedence)"

Yes

database credentials (Can be used with and without prefix for global configuration. "with prefix" takes precedence)

Yes

database credentials (Can be used with and without prefix for global configuration. "with prefix" takes precedence)

No

used for logging

NEXEED_GLOBAL_ENVIRONMENT_NAME_example

No

used for logging

NEXEED_GLOBAL_SYSTEM_NAME_example

No

used for logging

NEXEED_GLOBAL_APPLICATION_INSTANCE_ID_example

No

general log level

WARNING

No

log level for request based logging

ERROR

No

"list of tentant_ids for request based logging. use '[""tenantid_a"", ""tenantid_b""]' to print requested logs for a number of tenants. '[""tenant_visible""]' will print all request logs"

No

"list of user_ids for request based logging. use '[""userid_a"", ""userid_b""]' to print requested logs for a number of tenants"

No

Maximum number of concurrent processes. This will also determine how many AMQP messages will be prefetched by the framework

30

No

Enables or disables in memory cache for trained ai models

true

No

Number of models in cache at which cache will be pruned (most recently used models are kept) to get back to CACHE_DEFAULT_SIZE

50

No

Number of most recently used models in cache that will always be kept

70

No

Number of seconds a model may stay in cache until it will be reloaded from the database

600

No

The number of curves that are collected for and used by model training. NB: There is one additional curve needed to actually trigger the training.

2000

No

The minimum number of collected curves that need to be valid for training. A curve can be invalid for training in case of bad data quality.

1800

No

Maximum of concurrent trainings allowed

1

No

Process curves with an anomaly score above this prediction threshold will be labeled as an anomaly. NB: this is the internal anomaly score that will always be scaled to 100 for display in the UI.

31

No

Interval that processes with >=3 failed trainings need to wait before being eligible for another training attempt.

604800

No

Share of process curves to ignore for determining both the normal upper and normal lower length of a process series. The share of training curves between the normal lengths for that series is: 1 - 2 * DQ_LEN_QUANT. This means the normal_upper_length and normal_lower_length are set by the curve at the DQ_LEN_QUANT and 1-DQ_LEN_QUANT quantiles. Only the following values are supported: 0.01, 0.02, 0.03, 0.05, 0.075, 0.1

0.02

No

Factor that gets multiplied with the difference of both the normal upper and normal lower length with the mean of a process series to determine upper and lower thresholds for labeling a curve to be a data quality issue. (e.g. upper_threshold = DQ_LEN_FACTOR * (normal_upper_length - mean_length) )

4.2

No

Share of process curves to ignore for determining both the normal upper and normal lower standard deviation of a process series. The share of training curves between the normal standard deviations for that series is: 1 - 2 * DQ_STD_QUANT. This means the normal_upper_std and normal_lower_std are set by the curve at the DQ_STD_QUANT and 1-DQ_STD_QUANT quantiles. Only the following values are supported: 0.01, 0.02, 0.03, 0.05, 0.075, 0.1

0.02

No

Factor that gets multiplied with the difference of both the normal upper and normal lower standard deviation with the mean of a process series to determine upper and lower thresholds for labeling a curve to be a data quality issue. (e.g. upper_threshold = DQ_STD_FACTOR * (normal_upper_std - mean_std) )

5.0

No

Share of process curves to ignore for determining both the normal upper and normal lower mean of a process series. The share of training curves between the normal means for that series is: 1 - 2 * DQ_MEAN_QUANT. This means the normal_upper_mean and normal_lower_mean are set by the curve at the DQ_MEAN_QUANT and 1-DQ_MEAN_QUANT quantiles. Only the following values are supported: 0.01, 0.02, 0.03, 0.05, 0.075, 0.1

0.075

No

Factor that gets multiplied with the difference of both the normal upper and normal lower mean with the mean of a process series to determine upper and lower thresholds for labeling a curve to be a data quality issue. (e.g. upper_threshold = DQ_MEAN_FACTOR * (normal_upper_mean - mean_mean) )

15.1

No

Ratio by which the average data collection rate of active processes will be reduced to calculate an update to the INITIAL_MIN_DATA_RATE (new_min_data_rate = min(initial_min_data_rate, average_data_rate * (1 - dynamic_min_data_rate_decline)).

0.2

No

Number of quickest-collecting process types to keep during cleanup

4

No

Initial minimum messages to receive per day for a given process type (so that it will not be cleaned up)

350

No

time in days for a process to start collecting without the risk of beeing cleaned up

1

No

The number of days a process needs to have status "low_data_quality" before process can be reset by cleanup function

3

No

Interval for time-triggered cleanup routine for slots with low data rate or low data quality.

43200

No

Interval for time-triggered cleanup routine to ensure processes stuck in training are cleaned up.

14400

No

Interval for time-triggered cleanup routine to clean up unused process data from process data table.

43200

No

At this value the number of models in the cache will be pruned.

150

No

This is the number of most recently used models in the cache that will be kept after pruning.

130

No

Model cache will be invalidated after this number of seconds.

600

No

Process table cache will be invalidated after this number of seconds.

10

Yes

host on which RabbitMQ is listening

Yes

host on which RabbitMQ is listening

Yes

port on which RabbitMQ is listening

Yes

user which should be used to get access to RabbitMQ

Yes

In multi-tenant environments, use a separate vhost for each tenant/environment

Yes

Defines whether communication with RabbitMQ is TLS encrypted

true

No

Authentication method used when connecting to RabbitMQ. Can be one of 'EXTERNAL', 'AMQPLAIN', or 'PLAIN'

PLAIN

No

IP Interface Web server listens to (Currently only used for health endpoint)

0.0.0.0

No

Port used for HTTP socket (Currently only used for health endpoint)

8000

No

"Can be used to manually disable TLS for HTTP endpoint. If set to true, cert.pem and key.pem do not need to be mounted to container"

false

No

Port used for HTTPS Socket (Currently only used for health endpoint)

8443

Yes

"Certificate used by Server during TLS connections. Files need to be mounted to '/etc/ssl/server/{{SERVER_CERT_FILE}}'"

cert.pem

"( File needs to be mounted

"Keyfile used by server during TLS connections. Files need to be mounted to '/etc/ssl/server/{{SERVER_KEY_FILE}}'"

key.pem

No

used to toggle Elastic APM on / off

true

Yes

URL of Elastic instance for OpenTelemetry exporter

Yes

used for authentication with Elastic

Yes

used for authentication and exporting OpenTelemetry to Elastic using ELASTIC_APM_SECRET_TOKEN

No

Elastic APM Log-Level

INFO

No

Specifies which exporter is used for logs. Possible values are 'none', 'otlp'.

otlp

Yes

Necessary to reduce the load. Specifies the Sampler used to sample traces by the SDK.

"parentbased_always_on"

Yes

Necessary to reduce the load. Specifies arguments, if applicable, to the sampler defined in by OTEL_TRACES_SAMPLER. The specified value will only be used if OTEL_TRACES_SAMPLER is set.

none

No

"Will be used to provide internet access for data extraction."

None

Yes

"OPTION 1: TNS-connection string → (driver oracle+oracledb will be used) OPTION 2: jdbc-url → (python driver will be inferred from jdbc-schema ['oracle:thin:@//', 'sqlserver://', 'postgresql://']). jdbc-url is only accepted in this format: 'jdbc:sqlserver://mssql.db.svc.cluster.local:1433/ppm' OPTION 3: sqlalchemy compatible connection string (Can be used with and without prefix for global configuration. ""with prefix"" takes precedence)"

Yes

database credentials (Can be used with and without prefix for global configuration. "with prefix" takes precedence)

Yes

database credentials (Can be used with and without prefix for global configuration. "with prefix" takes precedence)

No

used for logging

NEXEED_GLOBAL_ENVIRONMENT_NAME_example

No

used for logging

NEXEED_GLOBAL_SYSTEM_NAME_example

No

ID of the Microservice. Used for logging and health details.

NEXEED_GLOBAL_APPLICATION_INSTANCE_ID_example

No

general log level

WARNING

No

log level for request based logging

ERROR

No

"list of tentant_ids for request based logging. use '[""tenantid_a"", ""tenantid_b""]' to print requested logs for a number of tenants. '[""tenant_visible""]' will print all request logs"

No

"list of user_ids for request based logging. use '[""userid_a"", ""userid_b""]' to print requested logs for a number of tenants"

No

Maximum number of concurrent processes. This will also determine how many AMQP messages will be prefetched by the framework

30

No

Enables or disables in memory cache for trained ai models

true

No

Number of models in cache at which cache will be pruned (most recently used models are kept) to get back to CACHE_DEFAULT_SIZE

50

No

Number of most recently used models in cache that will always be kept

70

No

Number of seconds a model may stay in cache until it will be reloaded from the database

600

No

Process curves with an anomaly score above this prediction threshold will be labeled as an anomaly. NB: this is the internal anomaly score that will always be scaled to 100 for display in the UI.

31

No

At this value the number of models in the cache will be pruned.

150

No

This is the number of most recently used models in the cache that will be kept after pruning.

130

No

Model cache will be invalidated after this number of seconds.

600

No

Process table cache will be invalidated after this number of seconds.

10

No

Timer used to reload configuration periodically (in seconds)

300

No

Path to configuration file used in data extraction feature (Data extraction disabled if config file cannot be found)

/app/config/dataextract.json

No

Timeout for http post to Azure event hub

10