Security
How Olimpia protects your traffic, your secrets, and your data, and which controls you have.
At the edge
- All web traffic goes through Olimpia CDN. Your apps get automatic HTTPS on
*.olimpia.ccand on your custom domains, with attack protection at the edge. The region’s servers only accept web traffic that comes from there. - Databases behind a proxy. External access to Postgres and Redis goes through a TCP proxy with anti-DDoS protection and per-IP limits on open and new connections. TLS is required: unencrypted connections are dropped.
- The region’s address is not public. No DNS record points directly to the servers where your resources run.
Your secrets
- Encrypted environment variables. Values are stored encrypted and are decrypted only to build and run your app.
- Per-bucket credentials. Each bucket has its own access key, with permissions only on that bucket. A leaked key can’t be used to touch another one.
- Rotatable credentials. You can rotate your Postgres or Redis password, or a bucket’s key, with one click. The old one stops working immediately.
- Tokens we don’t store. For your sessions and tokens, we store only a hash. Revoking one invalidates it immediately.
Your data
- Postgres without superuser access for you or the explorer. Your
appuser owns its database but is not a superuser. The console explorer connects with that same user, in read-only transactions. - Redis without admin commands. The
defaultuser can’t runCONFIG,ACL,SHUTDOWN, or other admin commands. - Backups under your control. You schedule Postgres backups to an Olimpia bucket or to your own S3, with the frequency and retention you choose.
- Resources capped per service. Every app, database, and Redis instance runs with its own memory and CPU ceiling, so one service cannot starve the others.
Your account
- Sign in with email and password or with GitHub.
- View and end your active sessions from the console.
- Sign-in attempts are rate-limited per IP and per account.
- Account → Activity shows what was done in your account from the console, the API, or your agents: what, when, and from where.
- Agents connect with OAuth 2.1 and PKCE or with personal tokens that you can revoke at any time.
Report a problem
If you find a vulnerability, email us at [email protected]. We respond quickly and keep you updated on the fix.