Question

App Platform Logs - Retrieve past run errors

Posted April 3, 2021 603 views
LoggingDigitalOcean App Platform

I have been running a node.js app using the App Platform since last year. I am having occasional errors from users submitting information which I have been able to see through log messages (through the control panel).

An app crashed about a week ago (25 March) and DO automatically re-deployed/re-started the app (same deployment id). Since the app was restarted, I have been unable to view the crash log (pre-25 March) or previous normal log activity, and it appears these have been overwritten with new normal log activity from post-25 march.

Note that my last deployment of this was on 1 March (same deployment id). I have been unable to view any logs on the same deployment ID between 1 March and 25 March.

How can I go back to view the crash log details from 25 March? Through the control panel and doctl I can only view post-25 March logs.

These answers are provided by our Community. If you find them useful, show some love by clicking the heart. If you run into issues leave a comment, or add your own answer to help others.

×
Submit an Answer
1 answer

Hi there,

I believe that the run logs are only stored to stdout and stderr so once the container is redeployed the logs are cleared out.

I’ve seen a similar question here, where a community member was looking for a way to retain the logs:

https://www.digitalocean.com/community/questions/how-can-app-platform-container-logs-be-retained

I believe that according to the accepted answer, this is not supported currently, but this is something that is being considered for the future.

There is also another suggestion for a community member saying that they’ve decided to use supervisord and store the logs into a file, then transported those log files to a remote syslog server via https://github.com/papertrail/remote_syslog2 in run_command before launching Docker Entrypoint.

Regards,
Bobby

  • Thanks Bobby that makes sense that the log is cleared when the container is reset. I think the utility of the logging would really increase if the run time errors were permanently logged, but perhaps that would dramatically increase the resources required.