Skip to main content

Version 1.18.1.6

Release Date: September 8, 2026 Release Type: Stable Previously published as: 1.18.11 (image tag staging-1.18.11)

The "No Payload" log level no longer stores message content at any point of a successful run, together with node log, monitoring display and startup fixes.

Backend Server​

New Features​

  • "No Payload" log level never stores the payload of a successful run: Until now a flow set to "No Payload" wrote the message body to the monitoring record at the start of the run and cleared it when the run succeeded, so the body was present for the whole run and stayed permanently if the message never finished. The body is now held only in a temporary buffer that is deleted when the run succeeds, and nothing is written to the permanent record. If the run fails, the buffered content is moved to the message record, so a failed message can still be diagnosed and retriggered from the start. This behaves the same on PostgreSQL and on Elasticsearch, and messages left in DELIVERING at this level are covered by the automatic recovery of stuck messages.

Bug Fixes​

  • Log detail of the node after Start is no longer empty: The log detail of the node directly after the Start node was blank, while every later node was recorded correctly. The content is now recorded for that node as well, on both PostgreSQL and Elasticsearch.
  • Delayer nodes appear in step logs: At the "All Steps" level a Delayer node produced no log detail record at all. It is now logged like the other nodes.
  • Messages with an unrecognized log level no longer stay in DELIVERING: A message whose flow carried a log level value the platform did not recognize was never finished and remained in DELIVERING. Such messages are now completed like any level other than "All Steps".
  • Node log attributes are recorded reliably: Attributes recorded with a node log could be lost when an internal HTTP request or response object failed to be converted to text. These internal objects are now discarded before conversion, and a value that still cannot be converted is kept as a marker instead of being dropped. The connector log level setting is also no longer written into the recorded attributes, so the same message is logged identically at every connector log level.

Frontend Server​

Bug Fixes​

  • Show Full Log and Connector Logs display the stored node payloads: Show Full Log and the Connector Logs tab blanked all node payloads of a message whose main record carried no payload, so a run logged at "All Steps" could be displayed as if it had been logged at "No Payload". Monitoring now always returns what is stored; the "No Payload" level is applied when the data is written, not when it is displayed.
  • A schema update that cannot be applied no longer prevents the administrator account from being created: During startup, a database constraint that could not be created rolled back the whole initialization, which on a fresh database left no users and caused every sign-in to be refused. The constraint is now attempted separately and reported as a warning when it cannot be applied. Failures of the startup initialization steps are also written to the log as errors instead of being ignored silently.

1.18.1.6 is a stable release. Previous: 1.18.1.5