<@U04S8PLJWMT> is it possible to retrieve from int...
# ask-ai
f
@kapa.ai is it possible to retrieve from internal postgres all the connection states?
k
Yes. Airbyte stores connection state in its internal Postgres database in the
state
table. [[State table](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#state)] You can retrieve all connection states with a query like:
Copy code
sql SELECT * FROM state;
For a Docker deployment, you can first connect to the DB, then run the query:
Copy code
bash docker exec -ti airbyte-db psql -U docker -d airbyte
and inside psql:
Copy code
sql SELECT * FROM state;
This returns one row per state record, including
connection_id
,
state
(JSONB with the state payload),
stream_name
,
namespace
, and
type
(
STREAM
,
GLOBAL
, or
LEGACY
). [[Incremental state](https://airbyte.com/tutorials/incremental-data-synchronization#step-2); [State table](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#state)]
f
@kapa.ai how is called the field that contains the json?
k
A message from kapa.ai
In the internal Postgres
state
table, the JSON field that contains the state payload is called
state
and its type is
JSONB
. [[State table](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#state)]
f
@kapa.ai I want to change the defaults for the following variables JOB_MAIN_CONTAINER_CPU_REQUEST REPLICATION_ORCHESTRATOR_CPU_REQUEST JOB_MAIN_CONTAINER_MEMORY_REQUEST REPLICATION_ORCHESTRATOR_MEMORY_REQUEST to "0.5" "0.5" "4Gi" "512Mi" respectively on airbyte deployed via abctl on a google VM that has 40Gi RAM and 10 CPU with an external postgres. How can I obtain it?
k
A message from kapa.ai
Those four settings are controlled via environment variables. Instance‑wide job resources are configured with: -
JOB_MAIN_CONTAINER_CPU_REQUEST
-
JOB_MAIN_CONTAINER_CPU_LIMIT
-
JOB_MAIN_CONTAINER_MEMORY_REQUEST
-
JOB_MAIN_CONTAINER_MEMORY_LIMIT
[[Configuring connector resources](https://docs.airbyte.com/platform/operator-guides/configuring-connector-resources#configuring-instance-wide-requirements)] For values/units: - Memory: e.g.
2g
/
2G
/
2048Mi
(Kubernetes-style); e.g.
JOB_MAIN_CONTAINER_MEMORY_REQUEST=2g
. [[CPU/Mem units](https://discuss.airbyte.io/t/1675)] - CPU: fraction of a single core, e.g.
0.5
for 50% of 1 core; e.g.
JOB_MAIN_CONTAINER_CPU_REQUEST=0.50
. [[CPU/Mem units](https://discuss.airbyte.io/t/1675)] However, the provided docs only describe the generic
JOB_MAIN_CONTAINER_*
knobs and do not document: - How
abctl
exposes these env vars (e.g., which config file/command is used to set them), - Nor how to set
REPLICATION_ORCHESTRATOR_CPU_REQUEST
/
REPLICATION_ORCHESTRATOR_MEMORY_REQUEST
specifically. They appear in an open Helm/abctl issue, but with no configuration instructions. [[Helm chart issue](https://github.com/airbytehq/airbyte/issues/65129)] Because of that, I cannot reliably give you the exact
abctl
commands or file locations to change these defaults on your GCE VM deployment. What is clear from the docs: 1. Instance‑wide defaults are set via environment variables in the deployment (for Kubernetes, typically via `values.yaml`; for other deployments, via whatever mechanism sets Airbyte’s env vars). [[Configuring connector resources](https://docs.airbyte.com/platform/operator-guides/configuring-connector-resources#configuring-instance-wide-requirements)] 2. The correct values for your target defaults would be:
Copy code
bash    JOB_MAIN_CONTAINER_CPU_REQUEST=0.5    JOB_MAIN_CONTAINER_MEMORY_REQUEST=4Gi
and, by analogy, you’d set:
Copy code
bash    REPLICATION_ORCHESTRATOR_CPU_REQUEST=0.5    REPLICATION_ORCHESTRATOR_MEMORY_REQUEST=512Mi
in the same environment‑variable mechanism, if your deployment exposes those vars. But since the knowledge sources do not explain how
abctl
wires these variables into your deployment or which file you should edit, I’m sorry, but there’s no information in the provided sources that describes the concrete steps to apply these defaults in an `abctl`‑managed install on a Google VM. To move forward, you would need `abctl`‑specific configuration documentation (not present in the supplied material) or to inspect how
abctl
generated your manifests and then inject these env vars there.
f
@kapa.ai where can I highlight an error on documentation? On this page https://docs.airbyte.com/platform/operator-guides/configuring-connector-resources#configuring-instance-wide-requirements is reported
Copy code
Connect to the database and run the following query with the image name replaced to find the connector definition id.
select * from actor_definition where actor_definition.docker_repository like '%<image-name>';
but field docker_repository doesn't exists on actor_definition
k
A message from kapa.ai
You’re right to flag this: the
actor_definition
table schema in the platform docs does not list a
docker_repository
column, only fields like
id
,
name
,
actor_type
,
resource_requirements
,
default_version_id
, etc. [[Actor definition](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#actor_definition)] Instead,
docker_repository
is documented on the
actor_definition_version
table. [[Actor definition version](https://docs.airbyte.com/platform/understanding-airbyte/database-data-catalog#actor_definition_version)] So the SQL example in the Configuring Connector Resources guide that says:
Copy code
sql select * from actor_definition where actor_definition.docker_repository like '%<image-name>';
is inconsistent with the published schema. [[Configuring connector resources](https://docs.airbyte.com/platform/operator-guides/configuring-connector-resources)] Regarding where to report this error: the provided knowledge sources do not mention any specific mechanism (feedback button, GitHub repo, or form) for reporting documentation issues, so I can’t point you to an exact location or process. The docs simply present the content without describing how to submit corrections. [[Configuring connector resources](https://docs.airbyte.com/platform/operator-guides/configuring-connector-resources)] Given that limitation in the sources, all I can confidently say is: - Your observation about
docker_repository
not existing on
actor_definition
is correct relative to the documented schema. - The documentation itself does not, in the provided excerpts, describe how or where to highlight such an error.