iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Need a server to try this on? Antihost offers Linux VPS plans with NVMe storage, full root access and a 14-day free trial.
To add RAM to a hosted VPS, resize its virtual-machine plan or replace it with a larger instance from a snapshot. You do not install a physical memory module, and attaching extra disk storage does not increase RAM. Most upgrades preserve your operating system and files, but the exact process, downtime, disk behavior, and IP address depend on the provider.
First confirm that memory is the constraint. A larger plan is the right fix for sustained memory pressure, swap exhaustion, or out-of-memory kills; it is not a cure for CPU saturation, slow storage, an undersized database cache, excessive application workers, or a memory leak.
Recommended Free Tools
What “adding RAM” means on a VPS
A VPS receives its memory allocation from the provider’s virtualization platform. The amount is part of the compute bundle, alongside virtual CPUs, storage and, often, a transfer allowance. There is no accessible DIMM slot and no supported way to plug a desktop memory stick or USB device into the virtual machine.
The normal operation is called a resize, change-plan upgrade, vertical scaling or, in some services, creating a replacement instance from a snapshot. The provider allocates a new memory limit and may also change vCPU count or disk size. A resize usually keeps the same operating system and application data, but it can require a shutdown, migration or replacement server.
#1 Best Overall
- Capacity: 16GB (2x 8GB Modules) | Type: DDR3 240-Pin | Speed: 1600MHz PC3-12800 / (PC3-12800E) | ECC Type: ECC-UDIMM (ECC Unbuffered DIMM) | Rank: 2Rx8 (Dual Rank x8) | Voltage: 1.35V
- Designed for ECC UDIMM Compatible Servers/Workstations (Rated Speeds & ECC Capabilities are CPU Dependent). Not Compatible with Desktops/Laptops.
- ECC Types can not be mixed | All installed modules must be ECC UDIMMs in order to function properly | A maximum of eight ranks per memory channel can be installed at once
- All A-Tech memory modules undergo stringent quality control testing to ensure dependable and reliable performance
- Backed by A-Tech's Limited Lifetime Warranty + Tech Support Team available to help before and after your purchase
You normally do not need to reinstall Linux or your applications. You do need a backup, a maintenance window and a plan for networking details such as public IP addresses and DNS.
Confirm that RAM is actually the bottleneck
Paying for a larger instance before measuring can leave the real problem untouched. Gather observations during the period when the service is slow or failing, not only immediately after a reboot.
Check host memory and swap
On a Linux VPS, start with:
free -h
vmstat 1 5free -h shows total, used, available and swap memory in readable units. The available figure is more useful than “free” alone because Linux uses otherwise idle RAM for cache. Repeatedly low available memory, rising swap use and non-zero swap-in or swap-out activity in vmstat indicate pressure. A single high reading after boot is not enough evidence.
Record the output with a timestamp and compare it with your monitoring history. Also note whether the pressure appears only during backups, deployments, traffic spikes or scheduled jobs. A predictable burst may be solved by scheduling or limits rather than a permanently larger plan.
Look for out-of-memory kills
When Linux cannot satisfy an allocation, the kernel’s out-of-memory (OOM) killer terminates a process. Check the system journal and kernel messages, for example:
journalctl -k --since "24 hours ago" | grep -i -E "out of memory|oom|killed process"
dmesg -T | grep -i -E "out of memory|oom|killed process"Use the command that your distribution permits; some systems restrict dmesg to root. Match any killed process to the time of an outage. On small Lightsail WordPress instances, AWS specifically documents a case where memory exhaustion can cause the Linux kernel to kill MariaDB. That symptom points to genuine memory pressure, but it still does not identify whether the database, PHP workers or another service consumed it.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Check limits outside the VPS
A process can be denied memory even when the host appears to have headroom. Inspect container limits (for example, a Docker or Kubernetes memory limit), systemd service limits, hosting-control-panel limits and per-user quotas. A container capped at 512 MB cannot use the host’s unused gigabytes until its own limit is raised.
Review application and database logs for allocation failures, worker restarts, killed requests and connection storms. Count application workers and estimate their peak resident memory. A web server with too many PHP, Python or Node workers can exhaust RAM while each worker appears healthy in isolation.
Separate RAM problems from other bottlenecks
If available memory remains comfortable and swap is unused, investigate the competing causes:
- CPU saturation: sustained high CPU or run-queue values call for more vCPU or more efficient code, not necessarily more RAM.
- Disk I/O: high wait time, a full volume or slow database storage points to storage performance or capacity.
- Database configuration: an oversized buffer or cache can consume all memory; an undersized cache can instead cause excessive disk reads.
- Worker limits and queues: reducing concurrency may stabilize a service more cheaply than scaling the instance.
- Memory leaks: a process whose resident memory grows continually needs a code, dependency or restart-policy fix.
Take a before-change snapshot of these measurements. It gives you a defensible comparison after the resize and helps you reverse a change if the symptom was misdiagnosed.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsHow the major VPS upgrade methods differ
The operation is provider-specific. This comparison focuses on the practical decisions that affect risk and downtime.
| Provider/service | How you obtain more RAM | Disk behavior | IP and data considerations |
|---|---|---|---|
| DigitalOcean Droplet | Resize in place; choose CPU and RAM only, or CPU, RAM and disk. | RAM-only choice avoids enlarging the disk; the combined option increases both. | Clean shutdown and a current snapshot are recommended before applying the resize. |
| Amazon Lightsail | Create a new instance from a snapshot and select a larger bundle. | The bundle fixes memory, vCPUs, storage and transfer together. | A replacement can receive a different public IPv4 address. Remap a static IPv4 address to preserve the endpoint. |
| Vultr Cloud Compute | Open the instance, choose Change Plan and upgrade to a higher plan. | The upgrade activates a plan with more vCPUs, RAM and storage. | Vultr states that instance data and the file system are not changed by the plan upgrade. |
Across providers, verify six details before approving the change: whether RAM-only resizing exists, whether the operation is in-place or snapshot-based, the expected reboot or migration interruption, preservation of IP and DNS, whether extra disk is mandatory, and how snapshots support rollback. Plan prices and availability vary by region, plan family and date; select the current option shown in your provider account rather than relying on an old figure.
Resize a DigitalOcean Droplet
DigitalOcean documents two resize paths: CPU and RAM only, or CPU, RAM and disk. The RAM-only path is useful when storage capacity is adequate. The combined path is appropriate when the workload also needs more local disk.
Rank #2
- A-Tech RAM Memory compatible for select DDR4 Server and Workstation systems only; (*WILL NOT WORK with Desktop or Laptop Computers/PCs*)
- 16GB RAM Kit (2 x 8GB Modules); DDR4 DIMM 288 Pin; Speeds up to 2933MHz PC4-23400 (PC4-2933Y)
- ECC Unbuffered UDIMM; 1Rx8 - Single Rank x8; JEDEC DDR4 standard 1.2V
- Improves system performance, workload capacity, and reduces bottlenecks by increasing memory (RAM) resources
- Note: This memory is ECC Unbuffered and cannot be mixed with different ECC types such as ECC Registered, ECC Load Reduced, or Non-ECC Unbuffered; (Memory compatibility can vary among different system models and their installed components; please verify compatibility and follow memory channel guidelines to ensure maximum performance)
1. Record the current state
Save the output of free -h, mounted filesystems, service status, firewall rules, public IP addresses and monitoring alerts. Confirm that backups are recent and that you can log in through an alternate administrative path if the primary service is unavailable.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →2. Shut down cleanly
DigitalOcean recommends performing the shutdown from the command line to reduce the chance of data corruption. Connect over SSH and run:
sudo shutdown -h nowWait for the SSH session to close and confirm in the control panel that the Droplet is powered off. Stop databases and write-heavy applications gracefully if your shutdown policy requires it.
3. Create a current snapshot
Take a snapshot after the clean shutdown and before changing the plan. Keep the snapshot until the upgraded system has passed application checks; it is your quickest rollback reference if the resize exposes a configuration or boot problem.
4. Apply the resize
- Open the Droplet in the DigitalOcean control panel.
- Open the resize action and choose either CPU and RAM only or CPU, RAM and disk.
- Select a plan with enough memory headroom for normal peaks and expected near-term growth.
- Review the disk choice, recurring charge shown in the account and any interruption notice.
- Apply the resize and wait for the operation to complete.
- Power the Droplet on if it does not start automatically.
A resize can involve a reboot or provider-side migration. Do not treat the control-panel progress message as proof that the application is healthy.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
5. Verify the guest and applications
Reconnect over SSH and run free -h. Confirm the reported total memory matches the selected plan, then check mounts, services, database connectivity, firewall rules, scheduled jobs and monitoring. Exercise a representative application request and inspect logs for failed starts or permission errors. Keep the snapshot until traffic has remained normal through a peak period.
Upsize an Amazon Lightsail instance
Lightsail uses fixed bundles containing memory, vCPUs, storage and a transfer allowance. You cannot independently add RAM to an existing bundle. The documented method is to create a new instance from a snapshot and choose a larger bundle.
1. Prepare the source instance
Measure memory and swap, record the public IPv4 address, list attached storage and verify that the latest backup or snapshot is complete. Lower DNS time-to-live in advance only if you control the zone and need to move a non-static address; do not assume the old address will remain valid.
2. Create a snapshot
In the Lightsail console, open the instance and create a snapshot. Wait until the snapshot status is complete. A snapshot captures the disk state at that point, so stop or quiesce write-heavy services first when your application’s consistency requirements demand it.
3. Create the larger replacement
- Open Snapshots in the Lightsail console.
- Select the completed snapshot.
- Choose Create new instance.
- Select a larger bundle with the required memory.
- Create the instance and wait for it to become available.
Because the replacement is a new instance, its default public IPv4 address changes when an instance is stopped and started. If clients, allowlists or DNS records depend on a stable address, attach or remap a static IPv4 address to the replacement before directing production traffic to it.
4. Cut over and validate
Confirm the operating system sees the larger memory allocation with free -h. Test the web application, database, background workers, cron jobs, firewall rules, TLS termination and outbound integrations. Update DNS or client configuration only after those checks. Keep the original instance stopped but available until the new one has survived a normal workload window; deleting it too early removes an easy fallback.
Change the plan on a Vultr Cloud Compute instance
Vultr’s documented Change Plan upgrade activates a plan with more vCPUs, RAM and storage and does not change the instance data or file system.
- Open the instance in the Vultr control panel.
- Select Settings.
- Choose Change Plan.
- Select a higher plan.
- Click Upgrade and confirm.
Expect a reboot or migration notice if the control panel displays one, and schedule the action during low traffic. After completion, verify memory with free -h, then test services, mounts, firewall policy, DNS and monitoring. “Data and file system unchanged” does not eliminate the need for a backup: a failed boot, application incompatibility or an incorrect assumption about the workload can still require rollback.
Rank #3
- A-Tech RAM Memory compatible for select DDR4 Server and Workstation systems only; (*WILL NOT WORK with Desktop or Laptop Computers/PCs*)
- 32GB RAM Kit (2 x 16GB Modules); DDR4 DIMM 288 Pin; Speeds up to 2400MHz PC4-19200 (PC4-2400T)
- ECC Unbuffered UDIMM; 2Rx8 - Dual Rank x8; JEDEC DDR4 standard 1.2V
- Improves system performance, workload capacity, and reduces bottlenecks by increasing memory (RAM) resources
- Note: This memory is ECC Unbuffered and cannot be mixed with different ECC types such as ECC Registered, ECC Load Reduced, or Non-ECC Unbuffered; (Memory compatibility can vary among different system models and their installed components; please verify compatibility and follow memory channel guidelines to ensure maximum performance)
Choose a plan with useful headroom
Do not select a size that merely raises total RAM above today’s average. Leave room for the operating system page cache, database maintenance, deployments, traffic bursts and the growth you can reasonably forecast. The correct amount depends on the resident memory of each service and its concurrency, not on a universal ratio.
Check whether the provider forces a larger disk or vCPU allocation with the memory increase. Extra vCPUs can help a CPU-bound workload, while a larger disk can increase recurring cost without solving memory pressure. Conversely, a plan with more RAM but too few vCPUs may leave a compute-bound application slow.
For a replacement workflow, include the operational cost of copying configuration, remapping an address, updating DNS, reissuing allowlists and validating secrets. For an in-place workflow, include the expected shutdown or migration interval and the possibility that a kernel or bootloader issue will require console access.
Free tools Windows power users keep installed
One-click scans. No signup required.
Swap can prevent crashes, but it is not extra RAM
A swap file gives the operating system room to move inactive memory pages to disk. AWS notes that this can reduce out-of-memory crashes on instances with less than about 1.5 GB of RAM. It does not increase the VPS plan’s allocated memory: disk is far slower than RAM, and sustained swapping can make an application feel sluggish.
AWS’s Lightsail guidance demonstrates a 650 MB swap file and persistence through /etc/fstab. Treat that as an example, not a universal size. Confirm available disk space, choose permissions that prevent other users from reading the file, and monitor swap activity after enabling it. A swap file is an emergency buffer while you resize or reduce memory demand, not a substitute for a properly sized plan.
When testing, watch latency and swap-in activity during the same workload that previously failed. If the service remains alive but response times deteriorate, reduce concurrency or resize the instance rather than allowing unbounded swapping.
Disk storage and RAM solve different problems
An attached block-storage volume increases persistent disk capacity. It does not add memory to the virtual machine. Use a volume when the filesystem is full, logs are consuming space or data must be separated from the boot disk. Resize the compute plan when the operating system or applications are exhausting RAM.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchDigitalOcean describes volumes as network-attached block storage that can be attached or resized. Network storage may also have different latency and throughput characteristics from local disk, so moving a database there is a performance decision, not a memory upgrade. Check filesystem mounts and application paths after any storage change.
Safe-change checklist
- Measure: record RAM, swap, CPU, disk usage, mounts, IP addresses, services, limits and the symptom you are fixing.
- Back up: take a fresh snapshot or verified backup before resizing, changing swap or altering worker limits.
- Schedule: use a low-traffic window and notify users; plan for a reboot, migration or replacement cutover.
- Select headroom: account for peak traffic, databases, background jobs and expected growth rather than only current averages.
- Protect connectivity: confirm console access, static-IP behavior, DNS TTLs, firewall allowlists and monitoring contacts.
- Validate: check
free -h, application health, database connectivity, mounts, firewall rules, DNS and alert delivery. - Observe: compare memory, swap, latency and OOM events with the pre-change baseline during a peak workload.
- Clean up deliberately: remove or adjust emergency swap and revise worker limits only after stable behavior is confirmed; retain the rollback snapshot for an agreed period.
Troubleshoot a resize that did not solve the problem
The guest still reports the old RAM amount
Confirm that the provider operation completed and that you rebooted when required. Check the instance or replacement identity, not a similarly named server. If the control panel shows the new plan but Linux does not, collect console output and contact provider support before making further changes.
The service fails after reboot
Use the provider’s web console or recovery mode. Inspect systemd status, boot logs, mounts, environment files, secrets and filesystem checks. A plan change should not alter application files, but a replacement from an older snapshot can legitimately omit changes made after that snapshot.
Users cannot connect
Check the public IP, static-IP attachment, DNS records, security groups or firewall rules and TLS certificates. Lightsail replacement instances commonly require an explicit static IPv4 remap; an old DNS record pointing to the previous default address will not follow automatically.
Memory usage remains high
Identify the largest processes and compare their growth over time. Revisit container limits, database buffers, worker counts and leak indicators. If the workload simply expanded to consume the available headroom, schedule another resize; if one process grows without bound, fix or restart that process instead of scaling indefinitely.
The VPS is alive but slower
Check whether the application is swapping, waiting on disk or constrained by CPU. More RAM can expose a storage or CPU bottleneck by allowing larger caches or more concurrent work. Tune one variable at a time and keep the pre-change measurements so each result is attributable.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
