ilert Alternative from Germany

What does per-user billing mean for an on-call rotation?

That costs grow with the number of people on call rather than the number of systems monitored. ilert Pro costs €19 per user per month with a three-user minimum; Scale costs €39. A six-person rotation therefore runs €114 per month and a twelve-person rotation €228 — whether ten systems sit behind it or a thousand.

The model has a clear logic. The value of an alerting platform accrues per person: each participant gets their own contact methods, their own quiet hours, their own cover arrangements and their own position in the escalation path. That is genuinely the work the product carries, and the billing reflects it.

The effect shows up wherever a rotation should be staffed broadly. And broad staffing is the healthy shape: a three-person rotation means being on call every third week and is the fastest route to burnout and attrition. An eight- or ten-person rotation is sustainable — and costs three to four times as much under per-user billing. The pricing structure therefore works against the team size you would choose for resilience reasons.

A second, frequently overlooked point concerns people at the edge of the rotation. The on-call lead who only appears at the final escalation tier. The support colleague who should be informed about incidents without acting on them. Under per-user billing each of these costs the full rate, and the obvious saving — not creating them — is precisely the one that is missed during an incident.

Against per-organisation billing it can fairly be argued that it merely redistributes the cost: teams with few people but many systems pay more there. That is true. The question before deciding is simply which of the two figures grows in your organisation — the number of people on call, or the number of monitored services.

Why is heartbeat monitoring alone not enough?

Because a heartbeat only reports what is still able to report. It is an inversion: a process calls an endpoint on a fixed schedule, and the absence of that call raises the alert. That is the right shape for background work — nightly backups, queues, cron jobs. For the question of whether a website is reachable from a customer in Vienna, it is the wrong one.

ilert offers 25 heartbeat monitors from the Pro tier and is, by its own standard, fully equipped: it positions itself as the alerting layer that receives events from Prometheus, Grafana, Zabbix, Datadog or a custom script. More than 100 integrations are the actual product core, and they are well built.

The consequence is nonetheless that detection has to happen elsewhere. Teams already running Prometheus with a blackbox exporter, or any other monitoring tool in house, already have that layer — for them ilert is exactly the missing piece. Teams that do not are buying the second half of the problem and must solve the first half separately.

The difference becomes clear with external checks. A heartbeat runs inside your own environment and says nothing about whether the data centre was taken off the network, whether a DNS record resolves incorrectly, or whether a certificate expires in four days. Checks from multiple locations answer exactly those questions — and the question of a majority among them is why a single failed request is not yet an incident.

Combining both is possible and common. It does mean two contracts, two interfaces and a translation between them that somebody maintains. For organisations with their own platform team that is a deliberate choice in favour of replaceability. For a team of five it is mainly one more thing that has not been tested when it matters.

How many SMS does an on-call rotation need per month?

Fewer than expected as long as alerting is configured cleanly — and considerably more as soon as it is not. A rule of thumb: every incident that escalates produces one message per tier per person addressed. Two tiers with two people each is four messages, plus acknowledgement and all-clear. A team with five serious incidents a month lands at 25 to 40 messages.

There is a real difference here, and it favours ilert. The Pro tier includes SMS and voice calls without a volume cap. LIVCK includes a monthly allowance — 25 credits on Team, 50 on Business — with packs purchased beyond that. For a team with frequent night-time pages this belongs in the calculation.

The number itself is deceptive, because it says less about system reliability than about the quality of the alerting rules. The most common reason for an exhausted allowance is not an unstable service but a check that fires on every timeout instead of waiting for a majority across several locations and several runs. Teams that consistently overrun their allowance usually have a detection problem rather than a volume problem.

Regardless of vendor, the same review is worth doing: how many of last quarter’s pages could have waited until morning? In most teams it is the majority. Alerting that distinguishes between "wake someone" and "appears in the report" reduces cost and attrition at the same time — and that distinction is a matter of configuration, not of vendor.