Everything about your database. On one screen.
Every HostBase project comes with built-in reports, logs, advisors, a data editor, Edge Functions hosting, backups and notifications — running on a real Supabase stack in the EU.
shop-production
Active🇵🇱 Warsaw · Medium · Postgres 17Requests / min
2,874
p95 latency
43 ms
Database CPU
31%
Connections
18 / 100
API traffic by service
Live requests
- 200GET/rest/v1/orders?select=*12 ms
- 200POST/functions/v1/stripe-webhook48 ms
- 200GET/storage/v1/object/public/avatars/…9 ms
- 200POST/auth/v1/token?grant_type=password61 ms
- 204PATCH/rest/v1/orders?id=eq.:id15 ms
Project dashboard preview — sample data
Reports
Traffic, latency and resources — over 1 h, 24 h or 7 days
Reports collect traffic from your project's API gateway and sample the database every minute. You see right away which service drives the load and whether the instance is big enough.
- Requests by service: REST, Auth, Storage, Realtime, Edge Functions
- 2xx / 4xx / 5xx codes, p50 and p95 latency, top paths
- Postgres CPU and memory against the instance limit
- Database size, connections, transactions per second and cache hit ratio
Requests 24 h
41.2k
p95 latency
43 ms
5xx errors
0.2%
Logs
Request logs and logs for every service in the stack
One place for the API request log plus Postgres, Auth, REST, Realtime, Storage, Edge Functions and gateway logs — with filters, search and a live tail.
- Filter by service, status and path
- Full-text search
- Refreshes every 5 seconds in live mode
- 14:02:11infoauthPOST /token grant_type=password 200 (61 ms)
- 14:02:12infofunctions[fn:stripe-webhook] POST /stripe-webhook
- 14:02:12infofunctions[Info] checkout.session.completed cs_test_…
- 14:02:14warnrestSlow query 812 ms: select * from orders where …
- 14:02:15errorstorageObject not found: avatars/user-12/old.png
- 14:02:17infodbcheckpoint complete: wrote 412 buffers (2.5%)
Advisors & performance
Problems surface on their own — with the fix
Advisors scan your schema and database statistics. Every finding comes with a description, the affected objects and SQL to copy. The performance view shows the slowest queries from pg_stat_statements.
- Tables without RLS, RLS without policies, SECURITY DEFINER in public
- Missing, unused and duplicate indexes
- Slowest and most frequent queries, cancel running queries
CriticalTable public.orders has no RLS
alter table public.orders enable row level security;
WarningForeign key without an index
create index on public.items (order_id);
Database
Table editor, SQL and Postgres tooling
Browse and edit data like a spreadsheet, run SQL and manage the database without installing a client.
- Filters, sorting, adding and deleting rows, new tables with RLS
- A SQL editor with snippets and grid results
- Indexes, extensions, roles, active connections and pg_cron jobs
- Auth users and a Storage file browser
| id | name | price | stock | active |
|---|---|---|---|---|
| 1 | Thermal mug Pro | 79.99 | 120 | true |
| 2 | City backpack XL | 249.99 | 32 | true |
| 3 | BT headphones Mini | 189.99 | 0 | false |
| 4 | LED desk lamp Eco | 59.99 | 210 | true |
select status, count(*) from orders group by 1; -- 4 rows · 18 ms
Edge Functions
Function hosting without a CLI or a server
Pick a template, write code in the editor and deploy with ⌘S. Secrets, JWT enforcement, an invocation tester, metrics and logs sit right next to the code.
- Templates: Stripe, Resend, AI chat with Claude, CORS, scheduled jobs
- Secrets with .env import and per-function verify_jwt
- Deploy via the Management API and MCP too
- stripe-webhookv141.8k / 24 h
- send-emailv91.6k / 24 h
- ai-chatv3341.8k / 24 h
Deno.serve(async (req) => {
return Response.json({ ok: true })
})Backups
Daily copies and one-click restore
Every active project gets a daily backup at 2:00 (last 7 kept) — on the Free plan too. Take a manual backup before a risky change, then restore the database or download the pg_dump file when needed.
- Daily and manual backups
- Restore with confirmation
- Download the .dump file
- daily-20261015-020000Daily · 18.4 MB · Ready
- manual-20261014-163512Manual · 18.1 MB · Ready
- daily-20261014-020000Daily · 17.9 MB · Ready
Notifications
Know before your database gets too heavy
The platform continuously analyzes the load on every database. When CPU, memory or connections approach the instance limit, you get an alert suggesting a larger size. When database or storage size grows, a higher plan.
- Alerts for CPU, memory, connections, 5xx errors and slow APIs
- Warnings at 80% and 100% of plan limits
- A bell in the dashboard and an in-project banner with an upgrade button
High database CPU load
Averaging 91% of allocated compute over the last 10 min (Nano).
Upgrade to MicroMCP & Management API
AI agents and scripts as first-class users
The MCP server and Management API give your agents and CI/CD pipelines the same capabilities as the dashboard — with a personal token and your role's permissions.
- Projects, keys, SQL (execute_sql)
- Edge Functions and secrets
- The service_role key never reaches the agent
$ claude mcp add --transport http hostbase \
https://host.ulewicz.space/api/mcpPlatform monitoring
We watch every database and the whole server
Our team sees live usage for the server and every customer database: CPU, memory, connections and traffic. That lets us react to overloads before you feel them in your app, and plan capacity ahead.
- Server usage and free capacity
- Each database's load against its instance
- Server and project alerts in a single view
Server CPU
37%
RAM
9.8 / 16 GB
Free slots
7
- shop-productionMediumCPU 34%RAM 61%
- crm-b2b-clientLargeCPU 12%RAM 44%
- ai-assistantSmallSuggest upgradeCPU 88%RAM 72%
- personal-blogNanoCPU 4%RAM 39%
Build in a weekend. Scale to millions.
Spin up a free project in 2 minutes. No credit card, no pausing after a week.