Proxmox Backup Server instead of Veeam? What the switch does to your lifecycle

Veeam Backup & Replication 12 shipped on 30/01/2023 and is supported until 01/02/2027. Proxmox Backup Server 3 shipped five months later and runs out on 31/08/2026. Five months apart at the start, five months apart at the end — only the other way round: 38 months of runway against 48, ten months of difference. That gap appears in no comparison table, because it is neither a feature nor a price.

First the thing that outweighs any date

A backup product is judged by restores, not by calendars. Two questions come before the lifecycle question, and they can end the discussion on their own. Can you restore, and have you tried?Not “the job reports green”, but a real restore of a real machine into a real network, timed, run by someone who did not build the setup.

And where does application consistency come from? Proxmox VE’s snapshot mode holds the filesystem still through the guest agent: if the agent is enabled and running, it calls guest-fsfreeze-freeze and guest-fsfreeze-thaw “to improve consistency”, as the backup chapter of the Proxmox VE documentation puts it. That is filesystem consistency. Whether your database or mail store returns in a state it considers clean is answered inside the guest — with a quiescing script, an in-guest agent, or a dump. And the Proxmox backup client is a Linux program: the documentation calls it statically linked and says it “should run on any modern x86-64 Linux system”. For a physical Windows server it is not the answer. If either question has no good answer, the rest of this article is about a decision that is not yours to make yet.

Two lifecycles side by side

Proxmox ties its dates to Debian. The Proxmox Backup Server FAQ lists every major version against its Debian base — version 3 on Debian 12, version 4 on Debian 13 — and puts its own end of support next to Debian’s. Veeam instead publishes two phases in its lifecycle table: after “End of Fix” a version gets no further updates, patches or hotfixes but stays supported and still receives security updates; only at “End of Support and Security Fix” does both stop.

Read the runway column honestly and the story collapses in both directions. Proxmox is not on a frantic treadmill — a good three years per major version is what Veeam managed with version 11 and is on course for again with version 13. Veeam 12 is the exception, not the rule, and even those 48 months are less than they look: by Veeam’s own table, version 12 entered End of Fix in November 2025, so the last fifteen months of that window are security fixes only. The one comparison that holds is the one at the top of this page, and it is the one you actually lived through if you moved in 2023: same year, ten months less runway.

The difference is not the length, it is the depth

A Veeam major upgrade is an application upgrade on a server whose operating system you do not touch. A Proxmox Backup Server major upgrade is a Debian distribution upgrade of the machine your backups live on. The official upgrade guide from 3 to 4 describes it without varnish: get to at least 3.4.2-1, run the pbs3to4 checklist tool, rewrite the package sources from bookworm to trixie, then apt dist-upgrade and reboot into a new kernel. It warns that network interface names can change, recommends out-of-band console access such as IPMI, and tells you to work over SSH in tmux or screen rather than the web console — because the web console is served by the thing you are upgrading.

The FAQ adds the sentence that should set your maintenance window: a major upgrade “must be carefully planned and tested and should never be started without having an off-site copy of the important backups”. A backup of the backup server, before touching the backup server.

And the dates cluster. The Proxmox VE FAQ gives Proxmox VE 8 an end of support of 2026-08 as well. Run the usual pairing and you get two Debian major upgrades on two classes of host in the same month — one on the hypervisors, one on the box you would restore from if the first goes wrong. That is not a tighter release cadence. That is a sequencing problem.

And free is not a property of exactly one of them

The usual reason to look at Proxmox Backup Server is licence cost, and below a certain size that argument carries less weight in both directions than it seems. Veeam publishes a Community Edition that is free for up to ten workloads and, since the Proxmox VE integration, covers Proxmox VE too — the page names Proxmox VE first among the platforms it protects. What it does not come with is technical support: the heading above the paid tier on that same page reads “Need Technical Support or Want To Protect More Than 10 Workloads?”. Under ten workloads, “free” is not the deciding argument it is usually presented as.

The other way round, Proxmox Backup Server is free software, but its own documentation separates the repositories: the enterprise repository is “the stable, recommended repository” and requires a subscription, while the no-subscription repository is meant for “testing and non-production use” and is explicitly “not recommended” on production servers “because these packages are not always heavily tested and validated”. Running it that way is entirely legitimate and very common. It only means you have decided against the vendor’s own production recommendation — and that decision belongs in the same document as the saving.

What goes in the calendar instead

The end date itself needs no software: it is in the table above, kept current on the product pages for Proxmox Backup Server and Veeam Backup & Replication, and on the end-of-life overview next to everything else that is running out.

What most calendars lack is not the deadline but the trigger. With Proxmox the clock on a cycle starts when its successor appears: PBS 4 shipped on 06/08/2025, and from that day PBS 3 had just under thirteen months left. The release is the event you can be notified about — patchletter emails you when a tracked product publishes a new version. On the day the next major version lands you know the clock on yours has started, instead of noticing later at a repository that has gone quiet. Free, no account, one-click unsubscribe.

The honest limitation

Proxmox has published no end-of-support date for version 4 — the FAQ says TBA. The expectation of roughly three years comes from the Debian 13 cycle, not from a promise, and a table showing “no date published” is telling the truth rather than filling a gap.

The table above also mixes precisions. Both vendors publish months, not days; the day dates come from our own dataset, which mirrors endoflife.date. Veeam’s own table puts version 12 in February 2023 rather than on 30/01/2023 and version 13 in November 2025 rather than on 03/09/2025, and Proxmox writes 2026-08 where the dataset writes 31/08/2026. It moves a runway by a month or two, not the comparison. Where a consequence hangs on the date — an audit, a contract, a change approval — read the vendor table.

None of this says which product restores better, and a support date is the cheapest criterion on the list. It is worth exactly this much: it tells you when a decision stops being optional. Nobody announces the day itself either — patchletter reports new releases, not end-of-support dates, and no alarm goes off on 31/08/2026. As of 19/08/2026 neither product had an entry in the CISA catalogue of known exploited vulnerabilities — what presses here is a calendar, not an attack.

Cycle dates from our own dataset as of 19/08/2026, cross-checked against the vendor lifecycle tables; runway rounded to the nearest whole month. Lifecycle, repository and upgrade statements from the Proxmox Backup Server FAQ and installation documentation, the Proxmox upgrade guide from 3 to 4, the Proxmox VE FAQ and its backup chapter, the Veeam product lifecycle table and the Veeam Community Edition page — all retrieved 19/08/2026. Vendor dates change; check them at the source before you plan around them.