I run network-intensive workloads and…
I run network-intensive workloads and used LayerStack's Hong Kong cloud servers. The hardware is genuinely good (AMD EPYC). The policy around it is not. My plan included a multi-TB monthly transfer quota. Whenever my sustained traffic reached roughly >100 Mbps — legitimate application traffic, well within the monthly quota — their automated system flagged it as "abnormal" and blocked the server's network. Not throttled. Blocked. The service just goes dead until their system decides to unblock it. I asked support directly, in writing, three times: what sustained rate (in Mbps) is actually allowed on my plan? Their replies only repeated undefined phrases like "the capacity supported by your subscribed plan" and "the allocated limit" — never once a number. Their own pricing table lists memory, vCPU, storage and transfer volume, but no port speed and no permitted rate. Do the math: a 5TB/month quota averages ~15 Mbps. Real traffic peaks far above its average, so if peaks get you blocked, the advertised quota is mathematically impossible to consume. You are paying for TBs you are not allowed to use. Support also confirmed in writing: "there are no refund options available" — including for time your server sits blocked by their own system. Judging by other reviews here going back years, this is not a glitch, it is the business model: advertise big quotas, block anyone who actually uses them, keep the money. Every competent provider handles heavy users by rate-limiting (traffic shaping) — the service slows down but keeps working. Choosing to hard-kill a customer's network instead, with a secret threshold, is a deliberate choice. If your workload is a small website that never exceeds a few dozen Mbps, the servers work fine. If you plan to actually use the bandwidth quota you are paying for, stay away.