I had the case that the logstash container (7.8.1) just stopped working and we did not nominitor it. What I would have liked: in the momen it crashed docker should have restarted it from the scratch.
Question: Would using the restart policy 'always' achieved this? Is there a way to test this behaviour?
Thank you for your inside!
While not a direct answer to your question, you could leverage the Logstash monitoring API to check it's health. For a time, I was having issues keeping pipelines up even though Logstash was running, so I wrote a powershell script to query the API, and if it came back with unwanted results, the script would restart Logstash. I've got it in a repo on github, I think if you are just looking for up/down status, hitting the server on tcp/9600 will give you what you need.