Available in: Project development and Site production. This guide covers application and runtime Logs. Chat and Agent conversation activity and execution evidence are separate from application log views.
Logs record output and events from running work. Use them to connect a visible problem with the operation, environment, and code that produced it.
Begin in the resource where the issue occurred: Project development or a Site's production operation. A healthy development preview does not establish that production has the same release or configuration.
Find a known development event
Ask the Project to write the fixed message preview-check to server output when a development route is requested. Visit the route once in the web preview, then open Logs in the same Project. Enter preview-check in Filter logs and open the matching entry to check its time. For a released route, inspect the corresponding Site’s Logs.
Investigate a failure
- Note what failed and approximately when it happened.
- Select the relevant resource, environment, and time range.
- Read the first meaningful error and the surrounding output.
- Connect it to the release, startup, request, or Job involved.
- Make one focused correction and verify the original workflow.
| Evidence | Possible next step |
|---|---|
| Dependency or build error | Review source and installation output. |
| External timeout | Check whether the external action completed before retrying. |
| Missing configuration | Check the owning environment's secure settings. |
| Request denied | Inspect application and platform authorization separately. |
Preserve useful context
When requesting help, include the relevant time, resource, release, and a small sanitized excerpt. Avoid posting tokens, credentials, private request bodies, or unrelated customer information.
Logs have bounded retention and collection coverage. Missing output is not proof that an operation never happened. For production exceptions, use Monitoring; for aggregate traffic, use Analytics.
Mark an issue resolved only after checking the user-visible behavior. An absence of new error lines alone is not a successful verification.
Example: a request fails after release
A form worked in development but fails on the Site. Note the time and active release, then inspect production Logs for the request. Check whether the error concerns validation, configuration, database access, or an external response.
Correct the identified cause and repeat the original workflow. If an external write may have completed, reconcile its destination before retrying. See Site operations.
Ask for a focused investigation
Investigate the failed export on this Site around 14:30 UTC. Correlate the active release and Job evidence, identify the first meaningful failure, and propose a correction before changing production.
Include the actual Site and time range. Use Plans when the correction involves consequential choices, and keep credentials or customer payloads out of copied excerpts.