Perimeter 81 incident
Some users experience issues logging-in to the platform
Perimeter 81 experienced a critical incident on February 17, 2022, lasting 40m. The incident has been resolved; the full update timeline is below.
Update timeline
- investigating Feb 17, 2022, 05:37 PM UTC
Some users reported issues when trying to login to the platform. This issue affects only some of our users and the team is working on identifying the root cause and apply a fix.
- identified Feb 17, 2022, 06:04 PM UTC
The team is currently working on further identifying the root cause of the issue in order to apply a fix. Current ETA for fixing the issue is 30 minutes.
- resolved Feb 17, 2022, 06:18 PM UTC
The team was able to identify the issue and apply a fix. All users should now be able to login to the platform.
- postmortem Feb 23, 2022, 02:31 PM UTC
**Overview** Due to incorrect deployment of the web client's parameters, the URL to our authentication broker pointed to the wrong region. Our deployment tools that pre-populate the parameters to all components configured our system with the incorrect value. This resulted in timed-out authentication requests which prevented users from logging in to the platform and agent. Logged-in users with a valid authentication token weren’t affected by this issue and were able to log in and work as usual. **Detection** The Perimeter 81 Customer Success team received several reports from different users and escalated the issue to the R&D team as a critical issue. **Root Cause Analysis** The team pushed some new configurations and environment variables to the Perimeter 81 repository for the upcoming releases. An automatic action triggered an old process to deploy our Front End version with the new variables. This resulted in an old deprecated module replacing a new module in production. The old module had an incorrect routing to the authentication broker which triggered the issue. Since the deployment was not planned, we didn’t run the post-deployment regression testing and therefore the old module and wrong routing were not detected by the team. **Resolution** The correct environment variables were pushed back to the production system and automatically deployed to all relevant components.