Fully managed dedicated servers for teams without a sysadmin, bundling OS patching, monitoring, backup retention and a hardened baseline configuration on top of leased hardware. Anvilcore includes a managed control panel, quarterly security reviews and 24/7 phone escalation on its higher tiers. Root access is available but voids parts of the management SLA.
Share your experience to help others choose the right provider.
I expected a Nessus export with a logo on it. What we got was eleven pages written by an engineer who had clearly looked at our actual stack, including a finding about an exposed admin route that our own pentest missed the previous year. Hardware has been unremarkable in the good sense - no failures in fourteen months. Would recommend for regulated small companies specifically.
Three-person agency, none of us wanted to own kernel patching. Anvilcore does it, tells us when it happened, and the monthly report is short enough that we actually read it. Their on-call picked up at 23:40 on a Sunday when our app started OOMing and stayed on until it was resolved. Cheaper than hiring someone, obviously, and better than us pretending we know what we are doing.
The actual managed work is competent. My problem is that I can never predict the invoice. Backup retention overage, out-of-scope engineering hours, a line item once for "emergency change window" that was just them applying a patch we asked for. None of these are huge individually but the variance month to month makes it hard to budget. Ask for the full rate card before you sign, they will send it but they do not volunteer it.
Read the SLA carefully. Taking root access - which we eventually needed for a legacy binary that wanted a custom kernel module - voids the parts of the management agreement covering OS and security patching. So you are then paying a managed premium for a server you are managing. They were transparent when I asked, its all in the contract, but it is not the deal I thought I was buying.
Signed at a quoted monthly figure. Month two came in at roughly 1.8x because of backup retention we had no idea we were exceeding - there is no alert, no threshold warning in the panel, nothing. They did credit half of it when I pushed, to be fair, and then adjusted our plan. But the fact that you can silently run up a bill on a managed product is a design choice and its a bad one.
I expected a Nessus export with a logo on it. What we got was eleven pages written by an engineer who had clearly looked at our actual stack, including a finding about an exposed admin route that our own pentest missed the previous year. Hardware has been unremarkable in the good sense - no failures in fourteen months. Would recommend for regulated small companies specifically.
Three-person agency, none of us wanted to own kernel patching. Anvilcore does it, tells us when it happened, and the monthly report is short enough that we actually read it. Their on-call picked up at 23:40 on a Sunday when our app started OOMing and stayed on until it was resolved. Cheaper than hiring someone, obviously, and better than us pretending we know what we are doing.
The actual managed work is competent. My problem is that I can never predict the invoice. Backup retention overage, out-of-scope engineering hours, a line item once for "emergency change window" that was just them applying a patch we asked for. None of these are huge individually but the variance month to month makes it hard to budget. Ask for the full rate card before you sign, they will send it but they do not volunteer it.
Read the SLA carefully. Taking root access - which we eventually needed for a legacy binary that wanted a custom kernel module - voids the parts of the management agreement covering OS and security patching. So you are then paying a managed premium for a server you are managing. They were transparent when I asked, its all in the contract, but it is not the deal I thought I was buying.
Signed at a quoted monthly figure. Month two came in at roughly 1.8x because of backup retention we had no idea we were exceeding - there is no alert, no threshold warning in the panel, nothing. They did credit half of it when I pushed, to be fair, and then adjusted our plan. But the fact that you can silently run up a bill on a managed product is a design choice and its a bad one.
Onboarding was sold as a white-glove migration with a four hour window. It took two days and we had two separate outages during business hours, the second because DNS TTLs had not been lowered beforehand - something the migration team should be doing in their sleep. The people who run the service day to day are clearly a different and better team than the ones who move you in. If I did it again I would migrate myself and only hand over afterwards.