On Monday 23 November 2020 at 09:20 CET the first customers in the NL3 hosting location reported slowness while working in TOPdesk. Over the course of the day, several similar reports were logged.
TOPdesk investigated the issue and noticed a high CPU usage on several application servers. A major incident was created to group the requests and inform all customers.
We investigated if a common root cause could be found, but we didn't find any. No Operating system updates, malfunctioning hardware, or changes in the TOPdesk product could explain the slowness customers were experiencing. The problem was only found on 1 hosting location, and was evenly distributed over several unrelated machines, while only 10% of the hosting location appeared to have been affected.
We did notice some additional activity in the Java Virtual Machine used to run TOPdesk, but couldn't pinpoint the root cause.
On Tuesday there were very few reports of slowness. We closed the major by the end of Tuesday as no common cause could be identified and all affected customers reported better performance.
Follow-up actions
We identified 3 points of improvement were additional tools and information could help investigate or prevent these issues in the future. One task was already in progress, and will be continued with a high priority. We’ve forwarded the other ideas to our Performance experts for validation.