FAQ¶

License¶

We just want to try it — what does it cost?¶

Community edition covers 25 hosts and one tenant. Free. Setup as normal, no license key needed.

What's missing in the Community edition?¶

Nothing. Every feature is included without a license key too — agent, collector, active collector, auto-discovery, dashboards, AI assistant, log collection, reports, status pages, custom roles + RBAC, SSO (OIDC and Active Directory/LDAP), audit log, backup. The Community edition is only limited to 25 hosts and one tenant; a license key raises these limits. Details: License.

How long does a license run?¶

Indefinitely — a license key does not expire, and its limits apply permanently. It includes an update period: you can install every version released up to that date; Admin → System → Licence shows it as "Updates included until". Nothing is locked afterwards; only newer versions need a renewal. Details: License.

Can I move a license to another instance?¶

Yes: enter the key on the new instance under Admin → System → Licence and choose Move here; the old instance loses the slot at its next online contact. Without internet, use offline activation. Details: Moving.

Tenants and multi-tenancy¶

Do I need multi-tenant for single-org use?¶

No. One tenant is enough — the Community edition has exactly one anyway.

Can a user be in multiple tenants?¶

Yes. Roles apply per tenant: an account can be admin in one tenant and read-only in another — assigned directly on the account, via a team or via a tenant group. Which role applies where and where it comes from is shown under "Access & teams" on the user. Details: Roles.

Does a customer user see other customer tenants?¶

No. Every access is restricted on the server to the tenants in which the account has permissions; an object of another tenant answers like one that does not exist.

Agents¶

My agent shows „outdated" — what to do?¶

On every config fetch (every 5 minutes by default) the agent checks whether a newer version is available, downloads it, verifies the signature and restarts. Restarting the service does not speed this up — the first check comes with the first regular config fetch afterwards.

Agent token lost — now what?¶

Manage host → Connection → "Renew token". The old token is invalid immediately; reinstall the agent on the machine with the new token.

How much CPU/RAM does the agent use?¶

Little — the agent is a single program without a runtime. It needs more with active log collection and many script checks.

Can the agent run on macOS?¶

No. The agent is available for Linux and Windows.

Collectors¶

Do I need a collector per location?¶

If locations are separated by VPN: yes, one collector per location.

What does the collector do when server is unreachable?¶

It tries to send each result batch three times and keeps results only briefly in memory — what was measured during a longer interruption does not arrive (unlike the agent, which buffers results on disk). As soon as the server is reachable again, the collector runs all checks of its devices immediately instead of waiting for the next interval.

Does a collector scale with the number of SNMP hosts?¶

Up to ~500 hosts per collector without issues. Beyond that, set up a second collector and split hosts.

Checks and profiles¶

Why are some check types grayed out?¶

A check has to fit the device:

Agent checks need an agent on the host — devices without an agent don't get them.
Network checks (ping, port, HTTP, SNMP …) need a target: IP address or DNS name.
Collector checks need a collector assigned to the host.

I want to change a check type — can't?¶

Right, a check's type is fixed after creation. Delete the check and create it again with the new type.

How do I get a modified builtin back?¶

In the profile list, on the check: "Reset to shipped state" — the entry only appears for shipped checks that were changed.

Can I return multiple values from a script?¶

Yes — a script with output format Nagios and perfdata: OK - text | metric1=v1 metric2=v2 metric3=v3. From two metrics on, the history shows each as its own curve; the check's value is the first one. Details: Scripts.

Dashboard¶

My dashboard is slow¶

Possible causes:

Widgets without narrowing over many devices → narrow the data selection in the widget settings (tenants, folders, profiles, tags, devices)
Very long time ranges → for old windows Vesana computes from condensed values; shorter ranges are faster
Variables set to "All" over very many devices → restrict the selection

In the dashboard's ⋯ menu choose Share … → tab Public link (owner only): pick tenants, optionally expiry and password, copy the link. Details: Dashboard → Public link. The built-in overview can only be shared once you have customized it.

Do public dashboards lose data on token rotation?¶

No — the dashboard stays unchanged. Revoking the link invalidates it immediately; a new link gets a new token and the old one stops working.

Alerts and notifications¶

How do I prevent alert storms during major outages?¶

Several tools that complement each other:

Alert grouping per tenant (Tenants → expand tenant → Alert grouping): one summary message instead of single mails; if several hosts fail at once, one "Hosts unreachable: …" message arrives.
Reachability check per host: if the host is gone, one host message arrives instead of one per check.
Dependencies: while a parent host is unreachable, alerts of dependent hosts are suppressed.
Start-up phase after an outage (Admin → Notifications → E-mail & Messages, default 30 min, at most 2 h): problems that only start with the restart arrive as one "Alert after restart" message.

There is no per-channel throttle. Details: Too many notifications.

Alert arrives too late¶

Two knobs: (1) confirmation on the check — "Confirmations" × "Confirmation interval (s)" (default 1 × 10 s). (2) The alert rule's waiting time per status ("Notify once the problem persists for"): a problem must persist that long before a notification fires. On agent hosts, detecting a silent agent adds 3 minutes. Details: Alert rules.

Will I get recovery notifications?¶

Yes by default. Can be switched off in the alert rule per channel and step ("Recovery" switch) — the channel itself filters nothing.

What about time periods („alert only Mon–Fri 8–18")?¶

Alert rules have no time windows. For recurring quiet times, a recurring downtime works (daily, weekly on chosen days, monthly) — it then suppresses all notifications of the covered checks, not just those of one rule. Details: Downtimes.

Phone¶

Is there an app for iPhone/Android?¶

No store app: Vesana is added to the home screen from the browser (Android via the install button, iPhone/iPad via "Share → Add to Home Screen"). Push works on both — on iOS only inside the installed app. Details: Vesana on your phone.

Push does not arrive¶

Push is enabled per device by the respective user (Settings → Push → "Push on this device"), not per channel. The test button on the push channel tells you why nothing arrived (no device enrolled, subscription expired, push service throttled). Without HTTPS no browser allows push. Once the browser prompt has been declined, the permission has to be reset in the browser settings — Vesana is not allowed to ask again. On a phone, the device must allow the browser or the installed app to run in the background (battery optimisation), otherwise the operating system holds notifications back until the app is opened.

Updates¶

An update failed — what now?¶

The updates page states the reason (e.g. not enough free disk space, portal unreachable, migration aborted), and so does the update history. Before starting, it warns when disk space is short. A migration that hits a short table lock held by the still-running services is retried by Vesana itself. The updater creates a backup before every update. Details: Updates.

Can I pin a specific version?¶

Not directly. Vesana only installs new versions if "Automatic updates" is switched on under Admin → System → Updates (default: off) — otherwise the installed version stays until you choose "Update now" yourself. Separately, security updates and bug fixes of the installed version (e.g. 1.902.0 → 1.902.1) are installed automatically by default; the switch for that is on the same page. Instances with a tester licence receive new versions at the latest after the licence's deadline.

Backup and restore¶

How often do backups run?¶

By default, a backup of the configuration (without measurement data) runs daily at 02:00 UTC. At most 5 backups, 90 days and 10 GB are kept — the strictest limit applies. Schedules (including full backups) and retention are set under Admin → System → Backup & DR. Details: Backup.

How do I get the `FIELD_ENCRYPTION_KEY` back if it's gone?¶

You don't. Encrypted DB fields are then lost. So: store the key in password manager + printed in a safe.

Can I export a single tenant?¶

Not directly — a backup always covers the whole instance. Individual device profiles can be exported and imported.

API¶

Does Vesana offer a read-replica endpoint?¶

No.

Can I batch PATCH operations?¶

Some tasks have bulk endpoints — for example editing checks in bulk, bulk ACK, or a downtime for several devices. Otherwise it is one call per object.

What's the rate limit?¶

Sign-in, 2FA confirmation and identity confirmation: 10 requests per minute per client address (password change: 5). The other endpoints have no fixed per-user limit.

AI¶

What happens to our data at Anthropic?¶

The request content goes to the Anthropic API. For privacy concerns, choose local Ollama as provider. If Vesana finds too little in the wiki, it also searches the web — this applies with Ollama too; in the chat the web search can be switched off. Details: Data protection.

How much does AI cost?¶

Cloud providers bill per request — set a spending limit there. Local Ollama only costs your own hardware.

Do I need internet for AI?¶

Only with cloud providers. For local Ollama: only for model download.

Data and retention¶

How long is my data kept?¶

It depends on the kind of data — Vesana has three classes:

Events (status changes, audit log, sent notifications, anomalies): unlimited. They are tiny and cannot be restored, so nothing is deleted here by default.
Measurement series: raw data 90 days, hourly values 395 days, daily values unlimited. A long-term trend is never lost, it only gets coarser.
Raw material (logs from monitored systems 30 days, diagnostic events 90 days, heartbeats 90 days): expires — but the essence is kept: from the logs the daily tally, from the heartbeats the outages.

Configurable under Admin → System → Resources → Storage.

Can I see what happened on a device two years ago?¶

Yes, with two caveats:

Status changes, ACKs, downtimes and notifications are kept indefinitely in the device history — that is where you see what happened.
Measurements are available that far back in daily resolution. The check history and dashboard widgets automatically switch to the condensed tier once the window is older than the raw retention; the curve gets coarser but does not break off.

This of course assumes the instance was already running back then.

Why does my graph look coarser for old time ranges?¶

Because it is computed from the condensed values there instead of the raw data. Up to the raw data limit (default 90 days) Vesana computes exactly and time-weighted; beyond that from hourly or daily values. This is deliberate: it preserves the history without the database growing without bound.

I need the raw data for longer — is that possible?¶

Yes. Under Admin → System → Resources → Storage you can set "keep" for metrics up to 3650 days, or to 0 for unlimited. Do the math with the displayed growth per day first — raw data is by far the largest item.

Other¶

Can I put Vesana behind a VPN/bastion?¶

Yes, all web endpoints are pure HTTPS — any reverse proxy / VPN topology in front.

Does Vesana do telemetry?¶

Every instance reports a small, fixed record to license.vesana.org once a day (instance pulse) and learns in return whether its updates are released. With a licence key entered, a short licence check-in runs every 6 hours as well (identifiers and check values only). Content (hosts, checks, values, names) never leaves the server automatically.

Is ESXi supported?¶

Yes, via the vSphere API: the ESXi profile from the Vesana library reads the host with a script, an account with the "Read-only" role is enough (no shell access needed). Details: Check types.

Can I run Vesana on Kubernetes?¶

Docker Compose is supported. Your own additions to the Compose file belong in a docker-compose.override.yml, see Own customisations.

Next¶

Troubleshooting — structured symptom list
System health