Is there a way to see if there were any failed search queries on the cluster? I see the CPU usage was around 90% but I’m not seeing a way to confirm if this actually started causing any issues for users or not. We’ve already on 8vCPU cluster so I’d happy hear what are some methods to mitigate high CPU usage

One way to identify user impact is to look at the latency chart.

For around an hour, we started receiving reports from users that search was not returning any results and then we turned Typesense off for mobile users (switched back to another search engine). I initially thought we might have a bug with search parameters causing it to return no results for some flows, but I saw the 90% CPU usage and though maybe we had issues on the engine side instead. Latency chart has a spike as well, but nothing crazy. I wouldn’t really expect to see few failures from these charts, we reached ~22 searches/sec so there’s plenty requests happening to even few failures out

Can you please DM me your cluster ID and the time frame in UTC. I can check and get back to you.