When something goes wrong

How we act

1. 1

   Detect

   Our status page and a second server watch the platform day and night. Reports from researchers, users and suppliers reach a person.

2. 2

   Contain

   First we stop the leak: access closed, keys replaced, sessions and tokens revoked. We keep the evidence, the logs included.

3. 3

   Assess

   We record what happened, which data it touches, for how many people and with what consequences. A breach we do not have to report goes into our register too.

4. 4

   Report to the regulator

   If there is a risk to your rights, we report it within 72 hours of discovery to the Dutch Data Protection Authority, the Autoriteit Persoonsgegevens (article 33 GDPR). What is still unknown then, we add later.

5. 5

   Inform you

   If the risk is high, we tell you directly, in plain language, with what you can do yourself (article 34 GDPR). Organisers for whom we process data hear it within 36 hours at the latest.

6. 6

   Fix and learn

   We fix the cause, add a test that should have caught it, and put below what happened and what we changed.

Public log

Data breaches, security incidents and larger outages, with what happened and what we changed. Without details others could abuse.

1. Outage17 October 2026Resolved on 17 October 2026

   Two minutes unreachable during a restart

   During a planned restart of the main server after updates the site could not be reached for about two minutes, because switching to the backup server was still done by hand. Since then the backup server takes over by itself in a read-only mode, and we never do planned maintenance without switching over first.

2. Outage16 October 2026Resolved on 16 October 2026

   The site showed an error page

   From about 01:00 to 10:45 partyppl.nl showed an error page, because the web processes got stuck. We fixed the cause. Since then our backup server checks the site from outside every minute, so an outage is reported at once, in the middle of the night too.

3. Outage2 October 2026Resolved on 2 October 2026

   Background jobs fell behind

   The background jobs, such as updating the agenda, ran hours behind because one processor core was fully busy. The site itself stayed up. We spread the jobs over the hour and lowered the load.