FAQs

Here are some frequently asked questions and common issues that may occur in the Kleene app, and some simple steps that you can take to try to resolve.

Q: Why is a transform running out of schedule?

A: This may be because of conflicting schedule setups. Check if the transform has upstream dependencies whilst simultaneously being scheduled - if so, the transform shouldn’t be scheduled because the upstream dependencies should trigger its run.

Go to transform settings, toggle the manual tables off and save. Open the transform again and see the input tables updated, should be dependency based now. Reload it all to see it refresh and then the transforms should be aligned.

Q: What is a ‘skipped’ transform?

A: If there is a sequence of transforms and one of them fails, all downstream transforms will not execute and will get “Skipped” status.

With a chain of transforms, if the first one runs and then some transform in the middle fails, the non-executed transforms after it, get the “Skipped” status.

Q: How to cancel a running warehouse query?

A: Run a SELECT to find the target query to stop -

Snowflake

SELECT * FROM TABLE(information_schema.query_history()) WHERE execution_status = 'RUNNING'

Redshift

SELECT * FROM stv_inflight;

Then run the Cancel command on the query ID -

Snowflake

SELECT SYSTEM$CANCEL_QUERY( '<query_id>' );

Redshift

CANCEL <insert pid here>

Q: What’s the difference between an ‘Execution started’ log and a ‘Starting task’ log?

A: ‘Execution started’ log means that the transform has been entered to the queue. When they will actually start executing is affected by how many other transforms are in the queue and it will be overtaken if there is some other transform upstream that the entered one depends upon.

’Starting task’ log means that the transform is executing in the warehouse.

Q: Why am I seeing an SQL compilation error?

A: Extract runs that return compilation errors can be caused by a variety of factors, but some common reasons may be down to syntax issues.

Double check the extract configurations by making sure the destination table entered, doesn’t have spaces in it or start with numbers (e.g. 0-9). Also check that column names are appropriate, try to avoid special characters but generally we can escape them during the ingestion process.

Q: What is the Kleene IP address?

A: Kleene has two outward facing IPs -

54.78.204.135

34.242.207.164

Q: What if my extract triggers a ‘Zero Date Prohibited’ error?

A: When this error occurs, it is commonly fixed by updating the MySQL source with the extra parameters: enabledTLSProtocols=TLSv1.2&zeroDateTimeBehavior=convertToNull

Include this in the ‘Extra Parameters’ field near the bottom of the source setup.

Q: What are the ‘SOAP fault details’?

A: If an error triggers saying there is “Invalid client data. Check the SOAP fault details for more information.” It may be because there is no complete data available, which occurs when a report is requested with incomplete data for all included campaigns. Only partial data is available.

This means that the extract is requesting a set of data that is not fully in existence yet. Data for the time period may not be ready to extract, i.e. attempting to ingest data for an entire day when the day is not over. This can be resolved by trying to run/schedule the extract at a different time.

Q: Why was the connection reset?

A: If an extract fails on a Session Connection Reset, this will be because the connection to the source has been disrupted/killed/restarted. Usually seen with database connectors e.g. Postgres. There is more than one cause for this error, but an easy one to rule out is if the extracts are bottlenecking the connection.

Check if the extract schedule is too congested causing many (7+) extracts to start at the same time and overload the connection. Simple test is to run one of the scheduled extracts manually (outside of normal schedule) to see if it completes without the rest of the extracts utilising bandwidth.

Q: How to check the 'stl_load_errors'?

A: If an extract fails on ERROR: Load into table '<tablename>' failed. Check 'stl_load_errors' system table for details. , there may be more information stored in the Redshift table.

Run a SELECT on the system table -

SELECT * FROM STL_LOAD_ERRORS;

Check the err_reason column to see what extra information was logged about the error.