I’ve created a custom app which starts successfully.
It takes a few minutes to startup, but in the middle of the process I get an error from the UI that says there was a failure starting the app. I can’t seem to find any correlating logs or explanations for where this comes from, or why the ui doesn’t recognize a successful startup.
At this time the toggle shows “Starting” with the spinning wheel. The app timeout (600 seconds) is much longer than it takes for the app to start (~200 seconds) and the startup is still running when the error occurs at ~150 seconds. Checking the “wait” box doesn’t seem to make a difference. I can follow the whole process in my own logs and also verify that the app is running after the nvwb UI startup process gives me the error and then eventually says it’s not running. I can’t find any trigger or explanation for why the error would be received.
Currently thinking there’s some very small concern or functionality here that I must be missing or overlooking. I think it might be related to my health check command, even though I don’t have any trouble with the command in the terminal.
Again, I get the error:
- before my startup script completes / before the server finishes starting
- for a service that eventually starts up healthy, as verified w/health check command
- while the UI still shows “Starting” next to the app toggle
FYI - spec.yaml:
apps:
- name: jockey-server
type: service
class: process
start_command: ./code/jockey-server/app.sh 2>&1 | tee jockey-start.log
health_check_command: '[ $(nc -z localhost 8123 >/dev/null 2>&1; echo $?) -eq 0 ]'
stop_command: docker compose -p jockey-server down
user_msg: ""
logfile_path: logs/jockey-server.log
timeout_seconds: 600
icon_url: ""```
**Please tick the appropriate box to help us categorize your post**
[] Bug or Error
[] Feature Request
[] Documentation Issue
[x] Other