Logs and commands
Inspect your service and run bounded diagnostic commands.
Events and logs
Service events explain lifecycle changes. Container logs show application output when the regional runtime can provide it. Logs are plain text; they are not a readiness guarantee or a replacement for application-level monitoring.
curl "$OIY_API_URL/api/services/$SERVICE_ID/logs" \
-H "Authorization: Bearer $OIY_API_KEY"Run a command
A service must be running before command execution. The command interface is bounded and intended for diagnostics, not as a durable job scheduler. The API accepts a command string up to 8,192 characters and returns output, an exit code, and whether output was truncated.
{ "command": "python --version" }Send this JSON to POST /api/services/{id}/exec with authentication. Runtime execution is capped at roughly 65 seconds; the control request has a 70-second timeout. Use an actual workload process and task activity protection for longer work.
Uncertain outcomes
If command execution returns COMMAND_UNCERTAIN, the command may have run. Check logs and service state before trying again, especially if it writes files or changes application state. Do not automatically replay commands after transport failures.
Avoid commands that print secrets. Treat application log content as untrusted data.