This whole thread is dancing around the actual question. What certs do you need? Legally, for OVHcloud? None. Their Kimsufi line assumes you're competent enough to not set fire to the datacentre. Their SoYouStart tier is the same. If you want their Managed Kubernetes or Load Balancer add-ons, that's different, but the bare metal is just keys and a prayer.
I've administered forty-seven OVHcloud boxes across three projects here in Limburg. Never uploaded a cert. Never been asked. The "certified admin" fantasy is something HR departments invented to filter CVs, not something OVHcloud enforces.
What you actually need: working knowledge of their API, because their control panel lies about provisioning status, and patience for their billing department which operates on French business hours regardless of where your servers live.
Actually the API thing is correct because I have automated the Provisioning workflow with Terraform and the OVHcloud Provider and without the API Credentials you are actually clicking through the WebPanel like it is 2003. Actually the French business hours are a real Factor because I once had a Billing dispute that took eleven Days to resolve because it spanned a Weekend and a BankHoliday.
But actually the original Question was about Certs and the Answer is still Zero because OVHcloud does not verify anything except your CreditCard.
Amir's post above confirms what I suspected but wanted to verify.
Nested clarification for future readers:
- OVHcloud bare metal: no cert gate - sub-item: API access is the practical barrier, not accreditation - sub-item: billing support window is Paris-centric, plan accordingly
- "Managed" tiers: still unclear if they audit admin qualifications - sub-item: their SLA documents do not mention RHCE, LPIC, or LFCS - sub-item: this is distinct from employment postings that list such certs
The salary band analysis I provided earlier remains relevant for hiring, but the thread has correctly separated "what OVHcloud requires" from "what employers imagine they require." Two different questions that kept colliding.
OVHcloud API is flaky from SEA. I keep a tmux session on a Singapore Hetzner box just to proxy calls when their auth endpoint times out. No cert needed for that either, just patience and a fallback IP.
The real filter here is whether your provider's control panel loads at all when you're on hotel wifi in Đà Lạt. OVHcloud's does not, consistently. Their mobile app is worse.
The API thing. Yes. OVHcloud's panel told me a dedicated server was "delivered" for six hours while it was still in PXE loop. Their API returned the same fiction. I only knew because I netbooted my own diagnostics and the MAC wasn't even live on the fabric yet.
No cert teaches you that. No cert teaches you to distrust the vendor's own status page. That's scar tissue.
Re: Contabo. Their panel is slower but I've never caught it outright lying. Just omitting. Different sin.
Someone above mentioned Nagios. I ran that stack for two years at a Melbourne shop. Moved to VictoriaMetrics and PagerDuty. The difference isn't the tool, it's whether whoever set it up understood cardinality before they pasted the config from a Medium post.
PXE loop fiction. Yes. I've seen that exact pattern twice. The second time I had a provisioning script that polled their API every ninety seconds and logged the lie. Sent it to their enterprise support. Got a form letter in French.
Re: monitoring. I run Grafana with Prometheus on a separate provider entirely. OVHcloud monitoring OVHcloud is circular reasoning. When they have a backbone event, your alerts die with the path. Separate your observability from your infrastructure. This is obvious but I keep meeting teams who haven't done it.
Separate your observability. Aye. Learned that one the hard way in 2014. OVHcloud Strasbourg had a power event, my monitoring was on the same ASN, everything went dark including the thing supposed to tell me it was dark. Found out from a player posting on our CS 1.6 forum that the server was down. That's your monitoring stack right there: a Scottish teenager telling you on Steam that the box is fucked.
These days I keep a £5/month Vultr instance in London just for status pages and external polls. Cheapest insurance I buy.
Following this thread closely because I'm building out a cost model for a non-profit here in Vancouver and the cert question keeps coming up from our board.
Sub-list of what I've actually verified:
- OVHcloud Canada Inc. (Montreal) does not list cert requirements in their service terms - Their "Advanced Support" add-on mentions "certified technicians" but refers to OVHcloud's own staff, not yours - Two local colo providers (not OVHcloud) asked for insurance certs, not IT certs
Still unclear: whether any Canadian privacy legislation (PIPEDA, BC's own) creates an indirect requirement through "reasonable safeguards" language. Has anyone here actually faced an audit where admin qualifications were examined?
PIPEDA audit. No. I've been through two Australian Privacy Principle reviews. They wanted policies, access logs, encryption at rest. Nobody asked what acronyms the person who configured it had. They asked for documentation. The certless ferry guy from earlier? His runbooks were better than most ISO shops I've seen.
The board wants certs because boards understand framed paper. They don't understand that a runbook that actually works at 3 AM is worth more. I've started translating "no cert required" into "proven operational documentation required" for non-technical stakeholders. Same reality, different vocabulary.
Post a reply
You need an account to reply.
Log in or
register to join the conversation.
