Today between 1230 and 1630 CET users of the TrustSource UI experienced a 504.
This has been caused by the ALBs not accepting the healthcheck anymore. The TrustSource version installed June 10 - and running since then - required service scaling. The new instances did not result in accepted health checks, so the ALB kept killing the new instances. Despite several trials and different approaches, we were not able to get any instance of this image to be accepted as healthy. The same image working perfectly in other environments. We are investigating this together with our provider.
To overcome the felt outage, we rolled back to the version prior to June 10. Another update can be expected shortly.
However, the API-instances were not affected from this scenario. Thus, the impact has been very limited to UI access. we will update this with our concluding analysis.