<@U04S8PLJWMT> I have the following part on connec...
# ask-ai
f
@kapa.ai I have the following part on conneciton state
Copy code
"[\"communications\",{\"server\":\"communications\"}]": "{\"ts_sec\":1764050457,\"file\":\"db05-slave.087107\",\"pos\":90877811,\"row\":1,\"server_id\":347503800,\"event\":3}"
what are row and event values?
k
The provided knowledge sources describe: - What a connection and connection state are conceptually (cursors used for incremental syncs). [[Core concepts](https://docs.airbyte.com/platform/using-airbyte/core-concepts); [Modify connection state](https://docs.airbyte.com/platform/cloud/managing-airbyte-cloud/manage-connection-state)] - How to view and edit connection state JSON in the UI, and that it contains “the most recent values for the global or stream‑level cursors.” [[Modify connection state](https://docs.airbyte.com/platform/cloud/managing-airbyte-cloud/manage-connection-state)] However, they do not document the detailed schema of the state payload for specific connectors (e.g., what
row
and
event
mean inside a MySQL/MariaDB binlog cursor such as:
Copy code
json "ts_sec": 1764050457, "file": "db05-slave.087107", "pos": 90877811, "row": 1, "server_id": 347503800, "event": 3
I’m sorry, but there’s no information in the provided Airbyte docs that explains the semantics of the
row
and
event
fields in this connector-specific state structure. The knowledge sources only state that connection state holds cursor values, not how individual fields inside that cursor are defined. [[Modify connection state](https://docs.airbyte.com/platform/cloud/managing-airbyte-cloud/manage-connection-state)] To get an authoritative answer, you’d need connector‑level documentation or source code, which are not included in the current knowledge set.
f
@kapa.ai I need to change on connection state ts_sec, row, server_id, event to restart from a specific point
k
A message from kapa.ai
The Airbyte documentation and forum threads you provided only describe connection state at a high level (as an opaque “cursor” used by the source) and how to edit that JSON, but they do not document the meaning of connector‑specific fields such as
ts_sec
,
row
,
server_id
, and
event
, nor how to safely change them to restart from an exact point. [[State & checkpointing](https://docs.airbyte.com/platform/understanding-airbyte/airbyte-protocol#state--checkpointing); [Manage connection state](https://docs.airbyte.com/platform/cloud/managing-airbyte-cloud/manage-connection-state)] Key points from the docs: - State is treated as a black box: only the source connector knows how to interpret it; the platform doesn’t make assumptions about its internal structure. [[State & checkpointing](https://docs.airbyte.com/platform/understanding-airbyte/airbyte-protocol#state--checkpointing)] - Manually editing state is supported (in Airbyte Cloud via the UI, self‑hosted via DB/API), but should be done with extreme care because it can break syncs and require a full historical resync. [[Manage connection state](https://docs.airbyte.com/platform/cloud/managing-airbyte-cloud/manage-connection-state); [Manually defined cursor](https://discuss.airbyte.io/t/828)] - Examples in the forums show updating JSON fields in
state
directly via SQL (e.g. changing a timestamp cursor), but they are specific to that connection’s structure and don’t explain event‑level cursor fields. [[Update cursor manually](https://discuss.airbyte.io/t/393); [Manually defined cursor](https://discuss.airbyte.io/t/828)] Because of this, I cannot tell you which exact values to set for
ts_sec
,
row
,
server_id
, or
event
, or how to compute them for “a specific point” in your log stream, based on the provided knowledge. What you can safely do, within what’s documented: 1. Locate and back up the current state - Use the UI (Cloud) under Connection → Settings → Advanced → Connection State or query the
state
table / state API on self‑hosted to retrieve the JSON. [[Manage connection state](https://docs.airbyte.com/platform/cloud/managing-airbyte-cloud/manage-connection-state)] - Save a copy of the original state before changing anything. 2. Only change fields you fully understand - The docs and examples confidently show editing simple timestamp cursors (e.g.
created_at
,
updated_at
) when you know the meaning and format. [[Update cursor manually](https://discuss.airbyte.io/t/393)] - For low‑level log fields (
ts_sec
,
row
,
server_id
,
event
), there is no official guidance in the sources you shared, so modifying them is not safe based on this documentation alone. 3. If you must restart from an earlier point Within the documented behavior, the only fully supported and safe approaches are: - Reset / full resync of the connection or stream (let the connector rebuild state from the beginning or a configured start‑date), or - Use a higher‑level cursor (like a date column) if the connector supports incremental sync on that field, and adjust that value in the state instead of internal log pointers. [[State & checkpointing](https://docs.airbyte.com/platform/understanding-airbyte/airbyte-protocol#state--checkpointing)] Given the lack of connector‑specific state schema in the provided material, I cannot give you a precise procedure to set
ts_sec
,
row
,
server_id
, and
event
for a particular restart position without risking breaking your sync. You would need either: - Connector‑specific docs/source code, or - Direct guidance from the Airbyte team for that connector and version, as the Cloud docs explicitly recommend for manual state edits. [[Manage connection state](https://docs.airbyte.com/platform/cloud/managing-airbyte-cloud/manage-connection-state)]
f
@kapa.ai can I set debezium to debug logging?
k
A message from kapa.ai
Yes, you can enable more verbose Debezium-related logging in Airbyte via environment variables. From the Helm
values.yaml
example in the GitHub issue, you can control Debezium and CDC log levels per connector using env vars such as: [[Java heapspace issue](https://github.com/airbytehq/airbyte/issues/51616)]
Copy code
yaml global:   env_vars:     CDC_LOG_LEVEL: DEBUG     DEBEZIUM_MAX_QUEUE_SIZE_IN_BYTES: 536870912     #DEBEZIUM_LOG_LEVEL: DEBUG     MSSQL_CDC_LOG_LEVEL: DEBUG     #DEBEZIUM_LOG_LEVEL: DEBUG
In particular: -
CDC_LOG_LEVEL: DEBUG
– enables debug logging for the generic CDC layer. -
MSSQL_CDC_LOG_LEVEL: DEBUG
– enables debug logging for the SQL Server CDC implementation. - There is also a commented
DEBEZIUM_LOG_LEVEL: DEBUG
in that example, indicating you can set Debezium-specific logging to
DEBUG
as well if your deployment/connector supports that variable. [[Java heapspace issue](https://github.com/airbytehq/airbyte/issues/51616)] Be aware that Airbyte’s platform docs recommend not running with DEBUG logs enabled in production/self-managed at scale because it can overload the
server
pod. [[Scaling Airbyte – DEBUG logs](https://docs.airbyte.com/platform/enterprise-setup/scaling-airbyte#debug-logs)] So, in practice: 1. Add or update the relevant env vars (e.g.
DEBEZIUM_LOG_LEVEL: DEBUG
or
CDC_LOG_LEVEL: DEBUG
) in your Airbyte deployment configuration (such as Helm
values.yaml
). 2. Redeploy Airbyte so the new log level takes effect. 3. Use this only for troubleshooting, and turn it off again for normal, at-scale operation.