Skip to content

PRP: PostgREST Request Processing -- Leveraging db_pre_request and request headers to create custom claims, generate advanced, up-to-date, scalable RLS policies, implement allow and deny lists, rate limiting, and much more.

Notifications You must be signed in to change notification settings

burggraf/postgrest-request-processing

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

19 Commits
 
 
 
 

Repository files navigation

PRP: PostgREST Request Processing

Leveraging db_pre_request and request headers to: create custom claims, generate advanced, up-to-date, scalable RLS policies, implement allow and deny lists, rate limiting, and much more.

Overview

PostgREST provides a setting called db_pre_request that accepts the name of a PostgreSQL function. This function is executed at the beginning of each http request and can set up data (in memory) for the rest of the request. Why is this so powerful and so important?

RLS (Row Level Security) policies are evaluated for every row in a query, so if you're retrieving 1000 rows the select policy for the table will be executed 1000 times. To make things scalable, it's advised that you store "custom claims", which are stored in the auth.users.raw_app_metadata database field and sent as part of the JWT token with each request. This puts those claims into memory, making them fast to access (and much more scalable than doing 1000 individual database lookups.)

The problem with this approach is that 1. it's pretty restricting (limited to storing the claims in a specific field in a specific table in a hidden schema), 2. claims are only read when a user logs in, so those claims can become stale if they're changed before a user logs out and then back in, 3. claims can be cumbersome to create and update.

Using db_pre_request solves those issues by letting you store claims data anywhere you like. Since the db_pre_request function is run once for every request, claims are always current. Creating and modifying claims is also fairly easy and very flexible. The one drawback is that the db_pre_request function does need to run one time for every http request so it's important to keep that function as efficient as possible. This function can be optimized, though, through features such as conditional processing.

This method is also more secure than standard claims because no claims are ever exposed to the user.

Features

  • Access the JWT token
    • user info: id, email, phone, database role, metadata
    • session id
    • app metadata (authentication provider, other available providers, custom metadata)
    • 2FA data (i.e. is the user currently using 2FA?)
  • Access the request headers
    • connection information
    • client type
    • request content information
    • use connection information for:
      • allow-listing and deny-listing all or portions of your application
      • rate limiting users
  • Create custom claims
    • Read related data (such as a table of users and roles for your application)
    • Use any custom logic
    • Use the claims in RLS (Row Level Security) policies
    • Claims are always current because they're created on each request
    • Claims are only processed once per request so they're efficient in RLS policies

Samples

Restricting access to your application with RBAC or ABAC

  • RBAC (Role Based Access Control): RBAC Demo

Multi-tenancy application

Allow-listing by IP Address

Rate Limiting

Gathering basic usage data: country, platform, user-agent

Conditionally setting claims

using custom headers

based on the current user

Reference

sample: req()->>'claim_name'

{
  "aal": "aal1",
  "amr": [
    {
      "method": "password",
      "timestamp": 9999999999
    }
  ],
  "aud": "authenticated",
  "exp": 9999999999,
  "sub": "<uuid>",
  "role": "authenticated",
  "email": "<email>",
  "phone": "",
  "headers": {
    "host": "localhost:3000",
    "accept": "*/*",
    "cf-ray": "xxxxxxxxxxxxxxxx-ABC",
    "origin": "<host_url>",
    "referer": "<host_url>",
    "cdn-loop": "cloudflare; subreqs=1",
    "priority": "u=1, i",
    "cf-ew-via": "15",
    "cf-worker": "supabase.co",
    "sec-ch-ua": "\"Not.A/Brand\";v=\"8\", \"Chromium\";v=\"114\", \"Google Chrome\";v=\"114\"",
    "x-real-ip": "<ip_address>",
    "cf-visitor": "{\"scheme\":\"https\"}",
    "connection": "keep-alive",
    "user-agent": "<user_agent>",
    "cf-ipcountry": "US",
    "content-type": "application/json",
    "authorization": "Bearer <token>",
    "x-client-info": "supabase-js/2.26.0",
    "x-consumer-id": "<id>",
    "content-length": "2",
    "sec-fetch-dest": "empty",
    "sec-fetch-mode": "cors",
    "sec-fetch-site": "cross-site",
    "accept-encoding": "gzip",
    "accept-language": "en-US,en;q=0.9",
    "content-profile": "public",
    "x-forwarded-for": "<ip_address>, <ip_address>",
    "cf-connecting-ip": "<ip_address>",
    "sec-ch-ua-mobile": "?0",
    "x-forwarded-host": "<ref>.supabase.co",
    "x-forwarded-path": "/rest/v1/rpc/req",
    "x-forwarded-port": "443",
    "x-forwarded-proto": "https",
    "sec-ch-ua-platform": "\"<platform>\"",
    "x-forwarded-prefix": "/rest/v1/",
    "x-consumer-username": "anon-key",
    "x-credential-identifier": "<uuid>"
  },
  "session_id": "<uuid>",
  "app_metadata": {
    "provider": "email",
    "mysetting": "AAA",
    "providers": [
      "email"
    ]
  },
  "user_metadata": {},
}

About

PRP: PostgREST Request Processing -- Leveraging db_pre_request and request headers to create custom claims, generate advanced, up-to-date, scalable RLS policies, implement allow and deny lists, rate limiting, and much more.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published