Pricing

Why there is no per-event price

Analytics vendors charge per event because they store your events. Infralytiqs does not store your events — your ClickHouse server does. So there is no volume for us to meter, and no incentive for us to make sampling sound reasonable.

The practical consequence is that your analytics bill becomes a storage bill you already understand. With ZSTD codecs and LowCardinality dimensions on the event table, that bill is usually smaller than people expect — and if it ever is not, you change one TTL clause instead of negotiating a contract.

Unlimited events

There is no event allowance to exceed, because we are not paying to store them. The only limit anywhere in the ingest path is 1000 events per HTTP request, which is a batching detail rather than a quota.

Every visualisation

All fourteen panel types, the full dimension catalogue, all seven aggregations, drill-down, saved reports and realtime panels. Nothing is held back for a paid tier.

SSO sign-in

Google, Microsoft or any generic OIDC provider. Identity is not an upsell, because making teams share a password is a worse outcome for everybody.

Your own ClickHouse

Single node or replicated cluster, ClickHouse Cloud or a container on your own hardware. Registering it is the first thing you do, not a migration you graduate to.

The AEM integration

The OSGi bundle for AEM author and publish tracking is part of the platform. Asset events, login events and scheduled DAM storage snapshots included.

No lock-in by construction

Walking away means revoking our access to your database. The rows stay where they always were, in a standard ClickHouse table you can already query. There is no export to request.

What your ClickHouse will actually cost

This is the number people want, so here is how to work it out rather than a figure we made up. Take your monthly event volume, multiply by the compressed bytes per row, and multiply by the months you intend to keep.

Compressed row size depends on your payload — how many custom dimensions you attach, and how long your URLs are. The honest answer is to measure it, and after a day of real traffic you can, on your own database:

-- What one event actually costs you, compressed and uncompressed.
SELECT
    sum(rows)                                                   AS rows,
    formatReadableSize(sum(bytes_on_disk))                      AS on_disk,
    formatReadableSize(sum(data_uncompressed_bytes))             AS uncompressed,
    round(sum(data_uncompressed_bytes) / sum(bytes_on_disk), 1)  AS compression_ratio,
    round(sum(bytes_on_disk) / sum(rows), 1)                     AS bytes_per_row
FROM system.parts
WHERE table = 'website_events' AND active;

-- Extrapolate from your own measured row size to a monthly footprint.
WITH
    (SELECT sum(bytes_on_disk) / sum(rows)
     FROM system.parts
     WHERE table = 'website_events' AND active) AS bytes_per_row,
    50000000 AS events_per_month,   -- put your own figure here
    24       AS months_retained
SELECT
    formatReadableSize(bytes_per_row * events_per_month)                   AS per_month,
    formatReadableSize(bytes_per_row * events_per_month * months_retained) AS at_full_retention;

-- Usually long page_url values or a sprawling custom_dimensions map.
-- Worth knowing before you add another dimension to every event.
SELECT
    name,
    formatReadableSize(sum(data_compressed_bytes))    AS compressed,
    formatReadableSize(sum(data_uncompressed_bytes))  AS uncompressed,
    round(sum(data_uncompressed_bytes)
          / nullIf(sum(data_compressed_bytes), 0), 1) AS ratio
FROM system.columns_parts
WHERE table = 'website_events' AND active
GROUP BY name
ORDER BY sum(data_compressed_bytes) DESC;

A container, for evaluation

One docker run gets you a working ClickHouse on your laptop. Perfectly adequate for wiring up the integration and deciding whether you like the product, and it costs nothing.

ClickHouse Cloud

Managed, with a free tier to start and usage-based pricing after that. You get an HTTPS endpoint and credentials, paste them into the connection form, and you are done. Billing stays between you and ClickHouse.

Self-hosted, replicated

Your own nodes behind a load balancer. Tick the cluster option and give it a cluster name, and the table is created ON CLUSTER with a ReplicatedMergeTree engine. Cheapest at volume if you already run infrastructure.

Metered SaaS analytics	Infralytiqs
Who stores your events	The vendor	You do
What scales your bill	Event volume, priced per million	Compressed storage, at your provider's rate
Sampling above a threshold	Common, and often silent	None. Nothing in the pipeline samples
Retention	A tier boundary	A `TTL` clause on your table
Raw data access	An export request, or an add-on	Any SQL client, whenever you like
Joining against your own tables	Rarely possible	It is one database — just join
Leaving	Export, then hope the schema is legible	Revoke our access. The rows never moved
Data residency	Wherever the vendor's region is	Wherever you put the server

Questions about cost

What is the catch?

You operate a database. That is a real cost in attention even when it is small in money, and if nobody on your team wants to own a ClickHouse server then a fully hosted product may genuinely suit you better. In exchange you get no event caps, no sampling, no retention cliff and no export negotiation.

Is there any paid plan?

The platform as described here is free to use. Paid engagements exist for support commitments, private deployment of the collector and console inside your own network, and help with migrations or custom report work — and those are conversations rather than a pricing table. Get in touch if you need one.

Do I need a ClickHouse server before I can try it?

You need one before events can be stored, but a Docker container counts and takes about a minute. The ClickHouse guide covers a container for evaluation and a hardened setup for production.

Can several sites share one ClickHouse server?

Yes. A tenant groups sites, and each site is a separate stream. Point them at one database and filter by site_id, or give each its own database on the same server. There is also a reserved all-sites segment for tenant-wide configuration.

What happens if my ClickHouse is unreachable?

Ingest for that configuration fails while it is down, so treat the database with the same care as any production dependency. Client-side batching gives you some tolerance for short outages, and the log explorer in the console is where you will see the failures.

Start without talking to anyone

How to Get Started Talk to Us