Measured, dated and described
Hostilla.pl evidence centre
We publish what can be verified: scheduled page checks, response times, anonymous real-user metrics, Lighthouse results, a controlled restore test and source links for eTOP and DataHouse facts.
Latest public snapshot17 Aug 2026, 21:09 +02:00
No customer data, internal paths or credentials are published.
Scheduled observation
Availability and response time
Window: 27 Jul 2026 - 17 Aug 2026. Results cover the public critical-page set checked by the daily monitor.
How to read this: This is a transparent sample from a defined measurement point. It is not a promise that every future request will have the same result.
Lighthouse mobile
Synthetic performance of key purchase paths
The latest mobile run uses the same monitor and records score, paint timings, blocking time and layout stability.
| Page | Score | FCP | LCP | TBT | CLS |
|---|---|---|---|---|---|
| Home page (PL) | 96 | 1,385 ms | 2,551 ms | 62 ms | 0.007 |
| Web hosting (PL) | 99 | 1,230 ms | 2,044 ms | 19 ms | 0.000 |
| Domain search (PL) | 100 | 1,065 ms | 1,659 ms | 0 ms | 0.000 |
| SSL certificates (PL) | 99 | 1,228 ms | 2,046 ms | 42 ms | 0.000 |
RUM / Web Vitals
What real browsers reported in the last 24 hours
Measurements are anonymous and contain performance values, route class and device class, without logins, emails or customer domains.
The current sample is small, so these values are preliminary and are not used as an SLA.
- LCP p75
- 1,104 ms16 samples
- INP p75
- 64 ms1 sample
- CLS p75
- 0.00316 samples
- TTFB p75
- 570 ms16 samples
Measured before and after
Home-page mobile LCP reduced by 46.3%
Five comparable Lighthouse mobile A/B runs on 29 July 2026. We publish the median, not the best single run.
Before4,753 msLCP, Lighthouse 77
After2,551 msLCP, Lighthouse 96
Restore evidence
Controlled application-file restore test
Backup artifacts were copied into an isolated recovery directory and their SHA-256 hashes were compared with the pre-change manifest.
Scope matters: this test confirms readability and integrity of the named application backup. It does not claim that every customer backup was restored in this run.
PASS15files verified
Verifiable company and infrastructure facts
Evidence is more than a speed score
The source pages describe the operator, data-centre layer, network and editorial verification process.
Operator
eTOP sp. z o.o.
Źródło / SourceData centre
DataHouse / Polska
Źródło / SourceOperator network
AS20853
Źródło / SourceTechnical review
eTOP technical editorial team
Źródło / SourceMethod and limitations
How these numbers are produced
- A scheduled monitor requests a fixed set of public commercial and trust pages.
- HTTP status and full response time are aggregated over the visible measurement window.
- PageSpeed/Lighthouse measures four key mobile purchase paths.
- Anonymous RUM is shown with its sample size and marked preliminary when the sample is too small.
- Customer data, credentials, internal paths and raw security events never enter the public snapshot.
Snapshot generated: 17 Aug 2026, 21:09 +02:00
FAQ
Questions about the published evidence
Is observed availability the Hostilla SLA?
No. It is the percentage of HTTP 200 responses in the published scheduled sample. Contractual terms remain in the service regulations.
Are the Lighthouse values selected from the best run?
No. The current table comes from the latest scheduled measurement. The before/after case study explicitly uses the median of five comparable runs.
Does RUM identify customers?
No. The public data contains aggregate performance values and sample sizes, without login, email, phone number or customer domain.
What exactly does the restore test prove?
It proves that the listed application backup files can be read into an isolated recovery directory and match the pre-change SHA-256 manifest. Its scope is stated next to the result.
Can an editor run an independent Hostilla test?
Yes. The independent test programme provides a transparent methodology and a contact path for a test account.
Verify Hostilla.pl independently
Use our published method, request a test account or ask for a source behind a specific claim.