4

We run container instance group daily (triggered by logic apps). The container basically connect to queue, process it and ends. Sometime, according the events log, the container gets killed, there is nothing in logs, except the last thing our app did (and it was not at the end of processing). I checked the resources, we are way below the limits. Also, this does not happen every time (the container is the same, we did not push new version for some time) so its really baffling me.

Any ideas where I could look for reasons? Thank you very much

user3337015
  • 111
  • 4
  • 10
  • Did you get anywhere? I have the same experience with a Rest API that does nothing at the moment, but gets killed for no reason after a few hours. Microsoft Support have not been helpful either, they have no clue what is happening either. – jbx Mar 01 '21 at 10:28
  • @jbx Hey, we also contacted Azure support and were told, that multiple customers are experiencing it and they are investigating the issue, so hopefully they will be able to fix it – user3337015 Mar 02 '21 at 12:34
  • Cheers for your response. I have been communicating with support for over a week and getting nowhere. They asked me to switch on Log Analytics but it still didn't log anything. – jbx Mar 02 '21 at 15:10
  • This question is not new, but this issue is happening when I try to spin up Windows container only. With Linux it runs fine. – Fabio Mendes Soares Mar 22 '23 at 19:24

3 Answers3

0

Maybe it is not killed if it logs only the last thing, maybe if the queue is empty there is nothing to execute? If the same container behaves differently from time to time then maybe some external condition changes from time to time..microsoft docs - Retrieve container logs and events in Azure Container Instances

Igor
  • 266
  • 1
  • 3
  • 13
  • Hi, thanks for your suggestion. I know the work is not finished, because once it is, it is logged, so I know when the work is done. Thanks for the link, I will check it out – user3337015 Feb 18 '21 at 20:44
0

One issue could be the restart policy - perhaps try different restart policy options and have a look at the the Microsoft's ACI troubleshooting page. It could be that the container instance is killed at some point after the task is completed.

Christopher Dunderdale
  • 99
  • 1
  • 7
0

I had the same problem, in my case, my docker-compose.yml has health-check and the helth-check was incorrect and Azure aci restart my container after each unhealthy status, I fix the health probe and it works now

redis:
image: bitnami/redis:latest
container_name: 'redis'
hostname: redis
restart: always
ports:
    - 6379:6379
networks:
  - pe_network
healthcheck:
  test: [ "CMD", "redis-cli", "--raw", "incr", "ping" ]
  interval: 30s
  timeout: 10s
  retries: 5
Saeed
  • 3,415
  • 3
  • 24
  • 41