Hi,
I’m experiencing severe performance instability on my service, even though the resource limits are set to 8 vCPU and 8GB RAM.

Here are the issues:

1. Random 5xx bursts with 20–30% error rate even under very low traffic (<100 requests/min).

2. p99 response times spiking to 10–15 seconds, while p50 stays around ~700ms.

3. CPU usage staying near 0% even during slowdowns, which suggests:

   • throttling,

   • host-level contention,

   • container restarts,

   • or I/O stalls.

4. Large latency spikes on static file serving, with sudden egress peaks (visible in Metrics → Public Network Traffic).

5. No Serverless mode enabled. The service should be always on.

6. No code-level bottlenecks: the same backend (PHP + MySQL) runs 10× faster on a shared Aruba hosting.

Could you please verify:

• if my container is being throttled at the host level,

• if the underlying VM is overloaded (“noisy neighbor”),

• if there are internal restarts (OOMKilled, exit 137/143),

• if the storage layer is experiencing I/O contention,

• or if the allocated resources are not actually being provided?

I need consistent performance at the level advertised by the 8 vCPU / 8GB limit.
Please investigate the host machine and container logs on your side.

Thanks.