Frequently asked questions

Register with an email address, confirm the verification code, then open Deploy in the console. Pick a region (Noida, Mumbai, Chennai, Singapore or US West), an operating system, a plan or a custom shape, and a billing cycle. Pay the invoice or choose hourly billing from your wallet and the server is cloned, configured and booted in about 60–120 seconds. The first-deployment guide walks through every screen.

No. Verification is by email and phone. Indian customers can top up the wallet with UPI or net banking through Razorpay from about ₹150; international customers use cards, PayPal or crypto. New accounts that complete their billing profile receive a starter credit under the free tier.

The one closest to your users. Noida serves Delhi NCR and north India, Mumbai the west and Singapore at about 55 ms, Chennai the south (Bengaluru at about 12 ms), Singapore South-East Asia, and US West the Americas. All five sit on AS152565 with the same DDoS filtering; Singapore and the US carry a small price factor shown on the pricing page.

Plans are fixed shapes — Nano (1 vCPU, 1 GB, 20 GB NVMe) up to 2XLarge (16 vCPU, 64 GB, 400 GB) — at a fixed price. Custom builds let you set 1–32 vCPU, 1–128 GB RAM and 20–2000 GB NVMe with sliders, priced per component ($1.50 per vCPU, $1.40 per GB RAM, $0.06 per GB NVMe per month) with no surcharge for unusual shapes.

Usually 60–120 seconds. A worker clones a verified template on a Compute Fabric node in your region, resizes the disk, boots the VM and uses the QEMU guest agent to write the IP, hostname and password before the welcome email leaves. You can watch each step on the server page or through GET /api/v1/jobs/{id}.

On the server page in the console (and through the API) — never in email. Root or Administrator credentials are shown once provisioning finishes; you can reset them at any time through the guest agent, and add an SSH key at deploy time on Linux.

Yes. Enable TOTP (Google Authenticator, Authy, 1Password and similar) under Account → Security; recovery codes are generated at the same time. Sessions are bound to your device and network, rotate automatically and can be revoked individually from the same page.

Use one of the recovery codes generated when you enabled 2FA. If you have none, email support@joy.services from the address on the account; identity is verified before 2FA is reset, which can take a business day.

Yes. Invite team members under Account → Security → Team; each person signs in with their own credentials and 2FA, access can be scoped and revoked at any time, and every action is written to the audit log under their name.

Not without your consent. We access VM contents only for support at your request, to investigate a suspected Terms violation, or where legally required — and every staff action on your account, including reading credentials or opening the console, is recorded in an audit log you can see under Account → Security.

Every account verifies email and phone. Identity verification (KYC) is requested in specific cases — reseller and enterprise accounts, large or flagged orders, opening outbound port 25, and marketplace sellers who want the verified badge. Status shows as none, pending, verified or rejected on your account, and the request always comes through a ticket, never by email link.

Email security@joy.services with steps to reproduce. We acknowledge within 48 hours, fix confirmed issues promptly and credit researchers publicly (with permission) and with wallet credit. Scope and rules are on the Trust Center.

Every price is set in USD and shown in your currency — detected from your country and changeable with the picker in the header. Invoices are issued in your account currency and show the USD base. Indian customers pay in INR through Razorpay; 18% GST is itemised on every invoice.

UPI, Indian cards, net banking and wallets through Razorpay; international cards and PayPal; Bitcoin and Lightning through BTCPay. Wallet top-ups start at $2 (about ₹150). Joy never sees or stores card numbers — PCI scope sits with the payment providers.

The wallet is a prepaid USD balance. Choose Hourly (wallet) at checkout and the server is charged every hour it exists, running or stopped — stopped servers still reserve their disk and IP — so destroy servers you are done with. Hourly is ideal for CI runners, load tests and staging; when the wallet reaches zero, hourly servers stop and are kept for 72 hours.

Hourly from the wallet, or monthly, quarterly, six-month and annual invoices with 0 / 5 / 8 / 15% discounts. The price you paid for a term is the price for that term; renewal changes are announced 30 days ahead in the Updates feed.

Invoices are due in 7 days. Reminders go out before the due date and at each step; a service is suspended 3 days after the due date with data intact and terminated 14 days after the due date, when data is deleted and IPs released. Paying before termination reactivates the service automatically.

Unused wallet top-ups are refundable within 7 days. Provisioned services are non-refundable except where Joy fails to deliver, in which case the pro-rated amount is credited or refunded. Full rules are in the Refund policy; returns and faulty hardware are covered by the Return and Replacement policies.

Yes. Add your company name, address and GSTIN under Billing → Profile and every invoice is generated with them, downloadable as PDF and available through GET /api/v1/invoices. PO-based billing is available for enterprise accounts.

Yes — prepaid codes redeemable for wallet credit in USD, bought through a Sales ticket and delivered by email within one business day. Redeem under Billing → Redeem gift card. Details are on the gift-cards page.

KVM virtual machines on Proxmox-based Compute Fabric nodes with host CPU passthrough, NVMe storage and a 1 Gbps port, in every Joy region. Plans include 1–20 TB of monthly transfer; a 10 Gbps port is available as an add-on.

Yes. More vCPU, RAM and disk are applied live where the hypervisor allows and pro-rated on the next invoice, from the server page. Disk downgrades require a reinstall. Self-service ceilings are 32 vCPU, 128 GB RAM and 2000 GB NVMe.

Yes — a noVNC console from the server page, served through Joy's own console host so no hypervisor address is exposed. Use it together with a reboot when a server does not respond, and include what you saw if you open a ticket.

Windows Server 2019, 2022 and 2025 are offered on any plan with at least 2 GB RAM (Micro and above) with a $4 per month licence add-on. RDP is enabled at first boot and the Administrator password is shown in the console. See the RDP guide.

Yes. A stopped server keeps its disk and IP, so hourly servers keep metering and monthly servers keep their term. Destroy a server you no longer need; take a snapshot first if you might want it back.

Not on shared compute (VPS). Mining is permitted on dedicated servers with prior written approval, as set out in the Acceptable Use Policy.

Ubuntu 22.04 LTS, Ubuntu 24.04 LTS, Windows Server 2019, 2022 and 2025. Templates are verified per region, so the deploy page and GET /api/v1/locations only offer what can actually be installed where you are deploying.

From the server page or POST /api/v1/instances/{id}/reinstall. The IP address and VM id stay the same, the disk is wiped and a new password is generated. Switching between Ubuntu and Windows is a reinstall; take a snapshot first if you might want to go back.

Not on VPS — templates are built and verified by Joy so the guest agent and networking work every time. Custom images and ISOs via virtual media are available on dedicated servers.

The QEMU guest agent lets the platform set your IP, hostname and password, reset credentials and read live stats. Removing it breaks password resets and network configuration on reinstall; leave it installed. Templates are rebuilt monthly with current patches and virtio drivers — see reinstall and OS templates.

Templates are rebuilt monthly with current updates, the virtio drivers and the guest agent. After first boot, updates inside the server are yours to manage; the hardening guide covers the first ten minutes.

Yes — one static IPv4 per server from the pool of its region, configured before the welcome email arrives. It survives reinstalls, upgrades and billing-cycle changes.

Up to 8 extra IPv4 addresses per server at deploy time or later, $2 per month each. Larger blocks and BYOIP are available for transit customers with an LOA and valid RPKI ROAs.

From the Network tab on the server page. PTR records propagate within about 15 minutes. The DNS and rDNS guide covers forward records too.

Closed on new accounts to protect our address space. Open a ticket describing your mail setup (SPF, DKIM, rDNS) and it is opened for accounts in good standing. Relaying through a provider on port 587 works immediately.

A 1 Gbps port and the plan's monthly transfer allowance (1–20 TB). A 10 Gbps port is a $15 per month add-on. Overage is billed at the regional rate shown on your invoice.

Joy's own autonomous system. It originates 13 IPv4 and 6 IPv6 prefixes, keeps 394 BGP peers and five upstreams, ranks fourth in India for known peers, and carries every Joy server, transit customer and CDN edge. Prefixes have RPKI ROAs and RPKI-invalid routes are dropped. Details on the AS152565 page.

AS152565 announces six IPv6 prefixes, peers over IPv6 and accepts IPv6 announcements from transit customers today. IPv6 addressing on VPS is not yet a self-service option — ask by ticket for your region.

Yes — always-on volumetric and L3–L7 filtering on every VPS, dedicated server and colocation port on AS152565, at no extra cost and with no traffic cap for protection. Attacks are detected by NeuroMesh NetFlow telemetry and scrubbed before they reach your port.

Volumetric floods, amplification (DNS, NTP, SSDP, memcached), SYN/ACK storms and L7 request floods are mitigated automatically. Custom rules for game protocols can be requested by ticket, and the CDN adds bot filtering and an optional managed WAF for HTTP origins.

Yes, with Virtual Connect: you announce your prefixes to AS152565 over BGP inside a GRE tunnel, attacks are scrubbed at our edge and clean traffic is delivered to your router anywhere. From $200 for 100 Mbps; above 1 Gbps it is priced as DDoS-protected transit from $700 per Gbps. See Peering & network services.

Transit customers tag a /32 or /128 with community 152565:666 and it is dropped at every edge. VPS customers open an Urgent ticket and the on-call engineer applies it within minutes. Mitigation events that affect a component appear on the status page.

No. A snapshot lives on the same NVMe as the server and protects you from your own mistakes — a bad upgrade, a broken config — not from hardware failure. For real backups use backup space or copy data off-node. The snapshots guide explains both.

From the server page or POST /api/v1/instances/{id}/snapshot; roll back or delete from the same place. Snapshots are free on every plan. Take one before every reinstall, resize or major upgrade.

Off-node capacity for scheduled full-VM backups, added in 10 GB steps at $0.02 per GB per month to any plan or custom build, at deploy time or later, and scheduled from the server page.

Platform databases are backed up hourly and replicated to a second region; customer server contents are not backed up unless you add backup space or run your own. You own your data and are responsible for backing it up, as the Terms set out.

Single-tenant Intel Xeon Gold / Platinum and AMD EPYC systems with ECC memory, NVMe (optional RAID 1/10), IPMI / iDRAC access over VPN and 1–100 G uplinks, racked in Mumbai and Noida with Singapore on request. Pricing is per component — CPU per core, RAM per GB, NVMe per GB plus a port — with setup waived on annual terms.

Quotes return within one business day and are valid for 14 days. In-stock configurations are delivered within 48 hours of payment; custom builds (large memory, GPUs, storage nodes) in two to four weeks. Credentials and IPMI access arrive by ticket, never by plain email.

Yes — per U or full rack in Tier III facilities with redundant power and cooling, metered power, cross-connects into the Joy fabric at $100 one-time, and remote hands included for basic tasks (reboots, cable swaps, media changes) and billed hourly for project work.

Failed components on dedicated servers are replaced within 4 hours of diagnosis, 24/7, and the physical host is monitored by Joy. Data on failed drives is not recoverable by Joy — keep backups.

Yes, NVIDIA L4, L40S and A100-class nodes on six-month and annual terms as custom builds. Open a Sales ticket with the model, memory and term you need.

Open a ticket from the console — the fastest route; replies also arrive by email and you can answer from your inbox, with a [JOY-…] reference keeping the thread together. If you cannot log in, email support@joy.services from the address on your account. Support is in English and Hindi around the clock; Bengali, Tamil and Telugu during Indian business hours.

First human response: Urgent (service down, attack) 15 minutes, High 1 hour and Normal 4 hours, all 24/7; Low (feature requests, feedback) 24 hours on business days. Targets are first response, not resolution; hardware or network problems get an update at least hourly until fixed.

Joy targets 99.99% monthly network availability per compute region. Service credits of 5%, 10% or 25% of the monthly fee apply below 99.99%, 99.9% and 99.0% respectively, requested by ticket within 30 days; scheduled maintenance announced 48 hours ahead (up to 4 hours a month) and beta services are excluded. The full text is in the Service Level Agreement.

Yes — /status, refreshed every 30 seconds with component health per region, incidents with updates, a history and a JSON feed at /status.json for your monitoring.

Platform, network and billing questions are always included. Work inside the OS — applications, patching, databases — is not managed by default; managed operations are available under Enterprise, and migrations from another provider are quoted per case (often free on annual plans).

Reply to it with the word ESCALATE and a manager is paged. For a network-wide emergency, reply to any ticket email with URGENT in the subject and the ticket is re-prioritised automatically.

Yes — REST API v1 at /api/v1 with Bearer tokens and scopes, covering quotes, deployments, power actions, reinstalls, password resets, snapshots, live stats, billing, invoices, tickets and status. The API reference has curl, PHP and JavaScript examples.

Under Account → API in the console. Tokens are shown once, stored hashed, carry scopes (read, orders.write, instances.write, billing.read or *), can have an expiry and an IP allow-list, and are revoked individually. The API quickstart deploys a server from curl in five calls.

Yes — 600 requests per minute per token (a 429 with Retry-After beyond that), 10 deployments per hour per account, 10 power actions per minute per instance and 3 password resets per 10 minutes. The public status feed is cached for 30 seconds. If a large fleet needs more, open a ticket.

POST /api/v1/quote returns monthly, hourly and total prices in your currency for any region, plan or custom shape and cycle — the same numbers the deploy page shows — so pipelines can budget before they create anything.

Create returns a job id; poll GET /api/v1/jobs/{id} for the steps (clone → configure → boot → agent → network → ready) and read credentials from the instance once it is running.

Your server's disk is on NVMe in the region you chose. Account and billing data is stored in India with encrypted backups in Singapore. Details are in the Privacy policy.

The Terms of Service, the SLA, Acceptable Use Policy and Marketplace Agreement, the Privacy policy and the Return, Refund and Replacement policies. Material changes are announced 30 days ahead in the Updates feed and by email.

We disclose customer data only on receipt of a valid legal request under the laws of the country where the data is held, and notify affected customers unless prohibited. Counts of requests are published in an annual transparency note in the Updates feed.

Spam and open relays, phishing and malware, scanning or brute force against third parties, launching attacks or BGP hijacking, illegal content, large-scale copyright infringement, mining on shared VPS, and abuse of the free tier through multiple accounts — the full list is in the Acceptable Use Policy. Reports go to abuse@joy.services.

PCI DSS scope is delegated to Razorpay, PayPal, the card processor and BTCPay — Joy never stores card numbers. Formal certification of the controls described on the Trust Center is in progress; until then that page, the Privacy policy and our answers to your questionnaire are the reference.

The laws of India, with the courts of New Delhi having exclusive jurisdiction after a good-faith attempt to resolve disputes through the help desk. Consumers in jurisdictions that grant non-excludable rights keep those rights.

Private Compute Fabric clusters, dedicated racks, BYOIP, DR replicas across regions, managed operations inside the OS and a named engineer who knows your setup — under one contract, one invoice with GST and PO-based billing. Start with a Sales ticket or the form on the Enterprise page.

Yes. Resellers and MSPs get a white-label console and local-currency billing for their customers, and gift-card codes may be resold by registered resellers. Reseller accounts complete KYC and agree to reseller terms; contact Sales to set one up.

Yes for enterprise customers — data-processing agreements, security questionnaires, architecture walkthroughs and NDAs for investors and partners are arranged through Sales.

Yes. Open a ticket with the source details; migrations are quoted per case and are often free for annual plans. Larger moves get an architecture review — attach a diagram to a Sales ticket.

Email peering@joy.services with your ASN, the exchanges or facilities and expected traffic — sessions usually turn up within a few business days under an open peering policy. IP transit is $500 per Gbps on the 95th percentile, IX ports from $150 per Gbps; details on the peering page.