Know It Broke Sign up

Security

Built like it holds your keysbecause it does.

To watch your systems, Know It Broke has to hold API keys, passwords and alert tokens. This page says exactly how they are kept, in plain words, with nothing left vague.

At a glance

The short version

Your secrets are encrypted. Your password cannot be read. Your card never comes near us. Nothing on the site watches you.

The rest of this page is the long version of those four sentences.

How it is kept
Secrets at restAES-256-GCM
Encryption keyKept apart from the database
PasswordsPBKDF2-SHA256, salted, 100,000 rounds
Two-step sign-inAuthenticator app (TOTP)
Single sign-onOpenID Connect, on every plan
Card detailsHeld by Stripe, never by us
Sign-in cookieHttpOnly, Secure, SameSite
Trackers and ad scriptsNone
Scripts, fonts or styles from anyone elseNone
HostingCloudflare's network

Your secrets

Encrypted before they are stored

Header values, request bodies, passwords typed by browser steps, webhook addresses, PagerDuty keys, mail server passwords and text message keys are all encrypted with AES-256-GCM before they are written down.

The key that unlocks them is not in the database. A copy of the database alone is not enough to read a single secret.

  • AES-256-GCM, a fresh random value for every secret
  • Key held apart from the data
  • Shown as dots once saved
  • Never sent back to your browser
  • Replace a secret without ever seeing the old one
  • Single sign-on on every plan, no extra charge
  • Password sign-in can be switched off
  • Two-step sign-in with an authenticator app
  • Each code works once
  • Ten one-use backup codes for a lost phone
  • Salted PBKDF2-SHA256, 100,000 rounds
  • Repeated wrong tries are slowed down
  • Reset links: one use, one hour

Your account

A password we cannot read, and a second lock

Your password is stored as a salted PBKDF2 hash, never as itself. Nobody here can look it up, and a leak of the database would not hand it over.

Switch on two-step sign-in and a stolen password is not enough: signing in also needs a six-digit code from the authenticator app on your phone. It works with Google Authenticator, Microsoft Authenticator, Authy, 1Password and any other app of that kind. No codes by text or email, which can be intercepted.

Or sign in through your own identity provider. Single sign-on works with Microsoft Entra ID, Okta, Google Workspace, Auth0 and anything else that speaks OpenID Connect, on every plan at no extra cost. You can switch the password off altogether, so the account opens only through your provider and the rules you set there.

Sign-in is slowed down after repeated wrong tries. Reset links work once and stop working after an hour, and with two-step sign-in on, a reset link alone does not open the account.

Your browser

No trackers. No scripts from anyone else.

The site runs only its own code. There is no analytics tag, no advert pixel, no chat widget and no script from another company that could read what you type.

A strict content security policy tells your browser to refuse anything else. The console cannot be loaded inside another site's frame, and every change you make carries a check that it came from our own page.

  • Strict content security policy
  • No analytics, adverts or chat widgets
  • One cookie, used only to keep you signed in
  • Cannot be framed by another site
  • Every change checked against forged requests
  • Checkout hosted by Stripe
  • No card data stored here
  • Signed messages, checked every time
  • Invoices and card changes on Stripe's portal
  • Cancel there in two clicks

Your card

We never see it

Payment happens on Stripe's own page. Your card number goes from your browser to Stripe and is never sent to Know It Broke, so it cannot be leaked from here.

Messages from Stripe about your subscription are signed. Each one is checked before it is believed, and an old or copied message is refused.

Your systems

Checks that say who they are

Every check names itself as KnowItBroke in its user agent and comes from Cloudflare's published address ranges. The Setup page lists them so your firewall can let them in and nothing else.

Add a secret header to any check and your server can tell a real check from somebody pretending. Know It Broke refuses to call private network addresses, so it cannot be pointed at the inside of anyone's network.

  • Published address ranges to allow
  • A clear name on every request
  • Your own secret header on any check
  • Private and internal addresses refused
  • Test runs are rate limited
  • Each account's data kept apart
  • Run history removed after 30 days
  • Delete everything yourself, at once
  • Never sold, never shared
  • Only three suppliers: Cloudflare, Stripe, Resend

Your data

Yours, kept apart, and gone when you say

Every node, result and alert route belongs to one account and is only ever read for that account. Run history is kept for 30 days and then removed.

Delete your account from the Setup page and everything goes at once: nodes, history, alert routes, secrets and your status page. Your data is not sold, shared or used for anything but running your checks.

Straight answers

Can you read my API keys?

They are stored encrypted and the console only ever shows dots. The service has to unlock a key at the moment it runs your check, because it must send it to your server. That is the only time it is used.

What happens to my secrets if I delete my account?

They are deleted with everything else, at once. Nothing is kept back.

Do you store card details?

No. Stripe takes the payment on its own page. Card numbers never reach Know It Broke.

Is there two-factor sign-in?

Yes. Switch on two-step sign-in in Setup and signing in needs your password and a six-digit code from an authenticator app such as Google Authenticator, Microsoft Authenticator, Authy or 1Password. You get ten one-use backup codes in case you lose your phone.

Why an authenticator app and not a code by text or email?

Texts can be redirected to another SIM and email accounts get broken into. A code made on your own phone never travels anywhere, so there is nothing to intercept.

Is there single sign-on?

Yes, on every plan, at no extra cost. Connect your own identity provider over OpenID Connect: Microsoft Entra ID, Okta, Google Workspace, Auth0, OneLogin and others. You give it an issuer address, a client ID and a client secret in Setup. A test sign-in has to pass before it is switched on, so a wrong setting cannot lock you out. An account is one sign-in today, so it covers that one person. SAML is not supported.

Can I switch the password off?

Yes. With single sign-on on, tick "Only allow single sign-on" and the password no longer opens the account. Who gets in, and with what second step, is then decided by your identity provider.

Do you hold SOC 2 or ISO 27001?

No. Know It Broke is a small product and has not been through those audits. This page says exactly what is in place so you can judge it for yourself.

Where does it run?

On Cloudflare's network. There are no servers of our own to patch, leave open or forget about.

How do I report a security problem?

Email hello@knowitbroke.dev. You will get an answer from the person who wrote the code.

Know it brokebefore your customers do.