Update timeline
After extensive monitoring, we are considering this incident fully resolved. We appreciate your patience and understanding while we investigated this complicated and nuanced infrastructure issue.
The root cause was determined to be related to the database when certain conditions coincided. Changes to the database capacity were implemented, which should address the root cause. We are continuing to monitor.
As of 10:56 AM ET the hot-standby API server pool was swapped in, resolving the issue with degraded performance. We are continuing to monitor.
As of 10:48 AM ET the API server pool has started to experience degraded performance again.
As of 10:34 ET we have implemented a fix. The new API server pool is working correctly and the service has returned to normal. Because this was the second time this issue arose we have three follow-up activities: 1) We are closely monitoring these servers and our engineering team is ready to immediately intervene if there is any performance issue. 2) We have provisioned hot-standby API servers so that if we need to failover it can be done faster and with less service disruption. 3) We are continuing to investigate the root cause and are collecting additional data. Once the root cause is identified we'll implement a fix to mitigate future service disruption. We apologize for the disruption that this incident has caused, and we remain committed to continuous improvements to maintain the confidence in Pearl's platform reliability.
As of 10:18 ET we have identified the issue, which was related to the API server pools. We're failing over to backup API servers. Updates will be posted as progress is made.
As of 10:05 AM ET some customers are experiencing the following errors when trying to access the Pearl platform. - "502 Bad Gateway" message - "500 Internal Server Error" message - Slow page loading This appears to be related to the API server pool issue from Friday. We are actively investigating this issue and are working to a quick resolution. Updates on our progress will be posted on our status page (https://pearl.statuspage.io/). Thank you for your patience.