Setting Up data ingestion for your ServiceNow® data¶

Archived records¶

The connector does not actively reflect the records archived in ServiceNow on Snowflake side for the ingested tables. Assuming that you archive inactive records older than a certain date, the following apply:

• Any record archived before the connector ingested it (for example, before the initial load of the table) will not be present in the table on Snowflake side at all.

• Any record archived after it was already ingested by the connector remains on the Snowflake side with no indication of archive action occurring.

• Any archived record restored for a table that is already operating in incremental updates mode will not be ingested on the Snowflake side unless that record is also modified afterwards (with its sys_updated_on value being updated to current time).

• An archived record restored during the initial load of the table may be ingested on Snowflake side depending on its ID in the sys_id column.

If you want to bring the table with an active archive rule up to date, you can reload the entire table but any record archived or restored after the reload is finished will follow the same principles listed above.

ServiceNow archive tables ar_[table_name] can be enabled for synchronization. However, the first incremental update that follows the initial load of such table are searched for records created/updated past the date the initial load of the archive table has started. Because neither sys_updated_on nor sys_created_on are modified when the record is archived, records archived after the initial load of the archive table up to a certain point in time are missing in it on the Snowflake side. For example, if you archive records older than a year, then any record archived for a year after the initial load of the archive table is not ingested to the archive table on the Snowflake side. The archived records that were restored or deleted by a destroy rule following initial load of an archive table is never removed from it on the Snowflake side.

Modifying Data Ingestion Using Snowsight¶

To modify the ServiceNow® tables to be ingested or the synchronization schedule for the tables, do the following:

1. Sign in to Snowsight as a user with the ACCOUNTADMIN role.

2. In the navigation menu, select Data Products » Apps.

3. Search for the Snowflake Connector for ServiceNow®V2 app, then select the tile for the connector.

4. In the page for the Snowflake Connector for ServiceNow®V2, select Data Sync.

5. Select Edit tables button to enter into the editing mode.

6. Modify the tables you want to ingest:

   1. Search for the table you want to ingest.

   2. Select the checkbox in the Status column from the left to the table you want to select or deselect.

   3. Under Sync Schedule, select how frequently you want to synchronize the table between Snowflake and ServiceNow®.

7. Select Update data sync.

Enabling or disabling the table synchronization¶

This section describes how to enable or disable the synchronization of a table in ServiceNow®. Synchronization enablement can be done with both default and custom configuration.

CALL ENABLE_TABLES(<tables_to_enable>);

CALL ENABLE_TABLES(['table1', 'table2', 'table3']);

CALL DISABLE_TABLES(<tables_to_disable>);

CALL DISABLE_TABLES(['table1', 'table2']);

CALL ENABLE_TABLE('<table_to_enable>', <table_config>);

CALL ENABLE_TABLE('<table_to_enable>', <table_config>);

CALL ENABLE_TABLE('u_table', { 'include_columns': ['sys_id', 'sys_updated_on'] });

CALL ENABLE_TABLE('u_table', { 'exclude_columns': ['col_1'] });

CALL ENABLE_TABLE('<table_to_enable>', <table_config>);

CALL ENABLE_TABLE('sys_audit', {
  'row_filter': 'tablename = "incident" AND fieldname = "cause" AND newvalue LIKE "privacy"'
});

CALL ENABLE_TABLE('incident', {
  'row_filter': 'active = "true" AND sys_created_by STARTSWITH "support" OR sys_created_by ENDSWITH "admin" AND category IN ("Network", "Cloud Management")'
});

CALL ENABLE_TABLE('incident', {
  'row_filter': 'incident_state IN ("1", "2", "3")', -- "New", "In Progress", "On Hold"
  'include_columns': ['incident_state', 'description']
});

Specifying the Synchronization Schedule¶

The Snowflake Connector for ServiceNow®V2 synchronizes data from all ServiceNow® tables to Snowflake on a specified schedule. By default, all of the tables are synchronized once every hour (1h).

To change the default synchronization schedule for all tables, call the CONFIGURE_DATA_INGESTION_SCHEDULE stored procedure with the following arguments:

CALL CONFIGURE_DATA_INGESTION_SCHEDULE(<schedule>);

Where:

schedule

Specifies the frequency of the synchronization. You can specify one of the following JSON values:

• { 'type': 'interval', 'value': '<interval_value>' }, where interval_value is one of the following string values:

  • '30m'

  • '1h'

  • '3h'

  • '6h'

  • '12h'

  • '1d'

• { 'type': 'custom', 'value': { 'hour': <hour>, 'day_of_week': '<day_of_week>' } }, where hour specifies the hour in UTC timezone at which the ingestion should start, and day_of_week specifies day of the week when the ingestion should be performed. It is possible to use special expressions as a day of week:

  • '*' to run the ingestion everyday.

  • '1-3' to run the ingestion from Monday to Wednesday.

  • '0,5,6' to run the ingestion on Friday, Saturday and Sunday.

  Possible values that can be used in the expression for day_of_week configuration are:

  • '0' - Sunday

  • '1' - Monday

  • '2' - Tuesday

  • '3' - Wednesday

  • '4' - Thursday

  • '5' - Friday

  • '6' - Saturday

  Other non-digit values like '5L' indicating the last Friday of a month or 'FRI-SUN' indicating the range from Friday to Sunday are not supported.

It’s possible to configure ingestion schedule for a specific table during its enablement. To enable a single table and set its ingestion schedule, call the ENABLE_TABLE stored procedure with the following arguments:

CALL ENABLE_TABLE('<table_name>', <table_config>);

Where:

table_name

Specifies a ServiceNow® table name to enable.

table_config

Object including schedule property, which specifies the configuration of the table synchronization. Check schedule of CONFIGURE_DATA_INGESTION_SCHEDULE stored procedure for details.

For example to enable ingestion of table table_1 and synchronize data every 3h call the following stored procedure:

CALL ENABLE_TABLE('table_1', { 'schedule': { 'type': 'interval', 'value': '3h' } });

The connector also allows you to specify a different schedule for each table that is enabled for synchronization. To change the synchronization schedule for a selected set of tables, call the CONFIGURE_TABLES_SCHEDULE stored procedure with the following arguments:

CALL CONFIGURE_TABLES_SCHEDULE(<table_names>, <schedule>);

Where:

table_names

Specifies an array of table names for which you want to configure the synchronization schedule.

schedule

Specifies the frequency of the synchronization. Check schedule of CONFIGURE_DATA_INGESTION_SCHEDULE stored procedure for details.

For example to ingest tables table_1 and table_2 each Saturday and Sunday at 11:00 PM UTC call the following stored procedure:

CALL CONFIGURE_TABLES_SCHEDULE(['table_1', 'table_2'], { 'type': 'custom', 'value': { 'hour': 23, 'day_of_week': '0,6' } });

Copy

By default the connector tries to start the ingestion in 3 hour time window from scheduled start time. If it is not possible to start the ingestion within that time frame, for example, when the connector is ingesting other tables, the current scheduled run is not executed. The connector attempts to run the ingestion at the next scheduled time frame. It is possible to change the duration of the time frame by calling CONFIGURE_CUSTOM_SCHEDULE_START_INGESTION_WINDOW store procedure:

CALL CONFIGURE_CUSTOM_SCHEDULE_START_INGESTION_WINDOW(<window_length>);

Copy

where window_length is length of the window in ISO 8601 duration format. The duration must be rounded to a whole hour, and must last for at least 1 hour. For example, value 'PT12H' specifies a window that lasts for 12 hours, and 'P2D' specifies a window that lasts for 2 days.

If you only enable tables with custom schedules, this configuration only affects time it takes to create and refresh flattened views for the configured tables. The flattened views are created in the first ingestion cycle after the following conditions are met:

• Ingestion of metadata tables is finished

• Ingestion of the configured table has started.

If email alerts are enabled, Snowflake recommends changing the alert frequency to Once per day when using custom scheduling.

Specifying the data range start time¶

By default, the Snowflake Connector for ServiceNow®V2 synchronizes all the records in the corresponding ServiceNow® tables. For the tables with: sys_updated_on or sys_created_on columns (from now on here called time columns) present, it is possible to restrict the range of synchronized data by setting a data range start time - i.e. lower bound for the corresponding time column value of the records.

With such a configuration the records with the corresponding time column value older than the data range start timestamp is not ingested. The corresponding time column used by this procedure is determined in the same way as for the incremental updates .

To change the data range start time value, call the CONFIGURE_TABLES_RANGE_START stored procedure with the following arguments:

CALL CONFIGURE_TABLES_RANGE_START(<table_names>, <range_start>);

Copy

Where:

table_names

Specifies an array of table names for which you want to configure the data range start time.

range_start

Timestamp specifying the data range start time in TIMESTAMP_TZ format or NULL to unset the current value.

• If the ingestion of the table has not been started yet, the data range start time value is taken into account with the first ingestion.

• If the ingestion of the table has already been started (e.g. a reload is in progress), the data range start time value is ignored and (another) reload of the table(s) is required to filter out the records with too old corresponding time column value.

It is therefore recommended to set the data range start time before starting the first ingestion of a table (hence also before enabling it).

For example, if tables table1 and table2 have the required time column(s), in order to set the data range start time to 2022-11-23 07:00:00 UTC for theses two tables, run the following command:

CALL CONFIGURE_TABLES_RANGE_START(['table1', 'table2'], TO_TIMESTAMP_TZ('2022-11-23 07:00:00 +00:00'));

Copy

Then:

• for table table1, for example, if ts ingestion has not been started yet, all the records with corresponding time column value before 2022-11-23 07:00:00 is not ingested.

• for table table2, for example, if its ingestion has already been started, the data range time start value is ignored in all data synchronizations until reloading this table. During the reload all the records with corresponding time column value before 2022-11-23 07:00:00 is not ingested.

It is also possible to unset the data range start time. For example, in order to unset it for table table1, run the following command:

CALL CONFIGURE_TABLES_RANGE_START(['table1'], NULL);

Copy

Again, if an ingestion of table table1 has already been started, reloading this table is required to ingest all the records back from ServiceNow®.

Reloading Data in a Table¶

To reload data in particular table, call the RELOAD_TABLE stored procedure:

CALL RELOAD_TABLE('<table_name>');

Where:

table_name

Specifies the name of the table to reload.

When you call the RELOAD_TABLE stored procedure, the connector performs the following example:

1. The connector suspends the original table for ingestion temporarily.

   Note

   While the table is being reloaded, you cannot re-enable the table for ingestion.

2. The connector creates a separate temporary table for ingestion.

3. The connector ingests the data into this new temporary table. This table is visible in the CONNECTOR_STATS view as a table named with a __tmp suffix).

   Note

   Every reload takes data range start time value into account, which can restrict which records are ingested.

4. After the data is ingested, the connector replaces the data in the original table with the data in the temporary table.

5. The connector deletes the temporary table.

6. The connector re-enables the original table for ingestion.

During this process, you can continue to query the existing data in the original table. However, changes to the data in the ServiceNow® table won’t be reflected in the Snowflake table until the ingestion process completes.

To avoid overloading your ServiceNow® instance, reload only one table at time.

Canceling Table Reload¶

To cancel the process of reloading the data in a table, use the CANCEL_RELOAD_TABLE stored procedure as shown in the following example:

CALL CANCEL_RELOAD_TABLE('<table_name>');

Where:

table_name

Specifies the name of the table whose reload you want to cancel.

When you cancel the reload, the connector drops all temporary objects created during the reload. The table is then available for ingestion as part of the normal synchronization schedule.

Configuring the Size of a Single Page Fetch for a Table¶

The connector fetches data from a table by dividing it into smaller chunks called pages. Each API request to ServiceNow® fetches one page. Due to limitations on the Snowflake and ServiceNow® REST APIs, the size of the response from the ServiceNow® API cannot exceed 10 MB and the response should be returned in less than one minute.

To account for this, the connector limits the number of rows fetched within a single API request. This limit is the page size.

The connector uses the following process to determine the page size:

1. Initially, the default page size is set to 10,000 rows.

2. If the fetch request fails during ingestion because the response size is exceeded, the page size is gradually decreased by 1000, 100, 10 and 1 until the request succeeds or the final page size is set to 1.

3. The successful page size is saved in the connector state and this value is used by subsequent requests.

The current page size for a table is available in the TABLES_STATE view. To see the page size, run the following command:

SELECT PAGE_SIZE FROM TABLES_STATE WHERE TABLE_NAME = '<table_name>';

Where:

table_name

Specifies the name of the ServiceNow® table being ingested.

The process the connector uses for determining the page size may lead to inefficiencies. This process only reduces the page size. It does not increase the page size. This can happen in situations where a table has a single large row that causes the page size to be set to a lower value.

To avoid this situation, you can manually set the page size by calling the RESET_PAGE_SIZE stored procedure as shown in the following example:

CALL RESET_PAGE_SIZE('<table_name>');

Where:

table_name

Specifies the name of the ServiceNow® table being ingested.

Ingestion run¶

Ingestion runs for a given table are triggered according to the configured schedule. A run downloads all the relevant rows divided into pages mentioned in the previous paragraph from the source table in a loop.

Initial load and updates

As soon as a page of data is fetched, it is inserted into the corresponding event log table. At this stage the newly fetched changes are not yet available in the sync table or through flattened views. When it is done the next request with updated criteria is issued as long as any data is returned. When the ingestion run is complete, and there is no more data to fetch in the source table, an asynchronous merge task is triggered, that takes all the changes from the event log inserted since the last merge and applies them to the sync table. When it is complete, the data becomes available in sync table and flattened views.

Truncate and load

In truncate and load mode a temporary table is created for each ingestion run. Each fetched page of rows is first inserted into this temporary table (this table exists in the internal connector schema and is not available to connector users). At this stage the newly fetched changes are not yet available in the sync table or through flattened views, they still show data fetched in the previous run. When the ingestion run is completed, and there is no more data available in the source table, data from the temporary table replaces existing data in the sync table. All the fetched rows are also added to the event log. At the end the temporary table is dropped.

Monitoring progress

To check the status of a current or past ingestion run, you can query the CONNECTOR_STATS view. It’s visible in the STATUS column. It’s set to DONE only if data was successfully fetched and all the changes were applied to the sync table. When the ingestion is running or the merge to the sync table/replace of rows in the sync table has not been completed yet, the status is RUNNING.