Earlier today, a number of customers reported slow response times when using LeanKit Kanban. When we investigated we found that we could replicate the issue intermittently but not consistently. The majority of response times for our test account were a little bit worse than average but not outside of normal parameters. But for a few accounts, we confirmed that response times were extremely slow. Investigation into the hosting infrastructure did not revel any obvious causes. Resource utilization metrics were above average for that time period but again not outside of normal parameters.
Although we didn’t have a clear root cause, we began to take standard corrective actions based on the most probable causes. Response times returned to normal based on our own testing and feedback from those customers who had initially reported the issue.
Even though the immediate issue is resolved, we are by no means done investigating. We are currently combing over all log information to uncover the root cause. However, this incident has already revealed some areas where we need to improve. Our current monitoring and alerting did not raise this issue quickly enough. We had to find out from customer. We know that’s not acceptable.
So, we are going to greatly expand our automated monitoring & alerting to include numerous test organizations instead of the one organization that we are using now. And we will expand the locations from which we monitor response time to provide better global coverage.
As we learn more, we will keep you posted. We appreciate your patience and sincerely apologize for any issues we may have caused for you.