Hey I have bunch of questions regarding Airbyte: 1. Is there a convenient way of changing workspaces on UI (some kind of list of workspaces or something like that)? 2. I have updated airbyte on GKE from 0.35.28-alpha to 0.35.46-alpha. I'm using GCP Cloud SQL as Database. I do not want purge the table and recreate as I have lots of connections already created. I have received this error trying to reach Airbyte UI error on server occurred:

SQL [select * from "public"."actor_definition" where "public"."actor_definition"."actor_type" = ?::"public"."actor_type"]; Error while reading field: "public"."actor_definition"."resource_requirements", at JDBC index: 15 at org.jooq_3.13.4.POSTGRES.debug(Unknown Source)

3. I was developing new connector and in logs I see error, but data is saved in destination:

2022-03-07 11:52:04 ERROR i.a.w.p.a.DefaultAirbyteStreamFactory(lambda$create$1):70 - Validation failed: null
2022-03-07 11:52:06 INFO i.a.v.j.JsonSchemaValidator(test):56 - JSON schema validation failed.

Hi @Artur Siepietowski, 1. Not at the moment, multiple workspace management is only enabled in our Cloud offer as of today. 2. Does these logs comes from the migration process running in the bootloader? 3. This is probably because the schema you defined for a stream does not match the input data. Double check your schema to make sure it corresponds to your source data.

1. Does these logs comes from the migration process running in the bootloader?

It comes from airbyte-server. I have managed to migrate to new version by 1. Exporting workspace, 2. Dropping airbyte database 3. Running all deployments and pods with new version 4. Importing workspace It worked for me but I think it is not a correct approach? Should bumping version 0.35.28-alpha to 0.35.46-alpha went seamlessly without dropping db?

1. This is probably because the schema you defined for a stream does not match the input data. Double check your schema to make sure it corresponds to your source data.

By

input data

you means HTTP Body from service or data produced from method:

parse_response

? What type schema should have if there could be 0 or more objects? shouldn't it be:

"type":["null", "object"]

?

By

input data

you means HTTP Body from service or data produced from method:

parse_response

?

Absolutely

What type schema should have if there could be 0 or more objects? shouldn't it be:

"type":["null", "object"]

?

Yes!