Memory Reboot not reflected in Metrics?

I have a service that keeps “failing due to using more than 512mb” of ram. But when I check the metrics or logs I do not see evidence of this. If I check activity it shows multiple restarts due to ram but again no evidence in metrics or the actual logs.

There is also no evidence of suddenly high traffic bursts.

Hi @jeregrine. Sometimes when memory usage spikes suddenly it’s not visible in the metrics graph, which only captures data once every couple of minutes. You might try temporarily upgrading to a larger plan to see if that fixes the issue. We only charge you for what you use, and billing is prorated by the second.