Postgres & MySQL Compatible App-scoped auto-access

Postgres & MySQL —natively wired into your app.

Define databases as components. If your service and database live in the same Celeris Application, access is granted automatically—standard drivers work out of the box.

Application: ecommerce
api Service
Binder Identity
orders-db Database
Driver ready (psql/mysql)

No static passwords. No manual grants.

Same application → instant connectivity

Celeris understands your application topology. When a service and database live in the same app, permissions bind automatically. Your standard psql/mysql drivers just work.

Application: orders-app
checkout-api
✓ Auto-bound
Application: orders-app
orders-db
connection.ts
// DATABASE_URL injected automatically
const db = new Pool({
  connectionString: process.env.DATABASE_URL
});

// psql/mysql drivers just work
await db.query('SELECT * FROM orders');

No secrets in repo. No manual grants. Identity injected at runtime.

Branch Databases

Fork data safely for preview environments

Create instant database branches from production, staging, or any environment. Isolated data for PR previews, performance testing, or migration rehearsals.

orders-db

Primary • Production

Healthy
Production HEAD

orders-db / branch: pr-1842

Created from production @ HEAD

TTL: 7 days Linked: PR #1842 Active

Branch Type

Common Use Cases

PR Previews

Isolated DB per pull request

Performance Testing

Load test without risk

Migration Rehearsal

Test schema changes safely

DEMO: Create branch for PR → connect service → run tests

Know what's slow, why, and how to fix it

This isn't just a database—it's a database with a cockpit. Real-time metrics, slow query insights, and actionable recommendations.

orders-db

Postgres 15 • Production • us-east-1

Healthy 5 active branches
QPS

1.2k

↑ 12% from avg

p95 Latency

18ms

Target: <50ms ✓

CPU / IO

24%

Healthy utilization

Cache Hit

99.2%

Excellent

Connections

42/100

Pool: 42% used

Lock Waits

0

No contention

Storage

124GB

of 500GB allocated

Slow Queries

3

Click to view details →

Connection Pooling

Built-in pooler optimizes connection reuse. Perfect for serverless workloads.

Backups + PITR

Continuous backups with point-in-time recovery. Restore to any second.

Read Scaling

Add read replicas to distribute traffic. Scale reads across regions.

Auto Maintenance

Scheduled maintenance windows with safe rolling upgrades. Zero downtime.

AI Schema Agent

Ask → Diagnose → Propose → Approve → Apply

The Celeris Agent understands your schema, query patterns, and performance characteristics. Ask questions, get actionable recommendations, and apply changes with approval workflows.

Ask the Agent

Try these:

AI Response Ready

Response condensed into an approval-ready change

Query Plan Analysis

Seq Scan
2.1s
Filter
Bottleneck
Sort
0.2s
Plain English

The query is doing a full table scan on 1.2M rows because there's no index on the status column. Only 847 rows match the filter, but the database has to read everything first.

Ready to apply?
DEMO: Ask → diagnose → index suggestion → approve → apply

No standing access. Just-in-time.

Time-bound, reason-bound, approver-gated database access. Automatically expires and revokes. Full audit trail on every operation.

How JIT Access Works

1

Request Access

Developer requests access with reason, duration, and scope. Link a ticket for compliance.

2

Approval Required

Designated approvers review and approve. Supports 1-of-N or 2-of-N thresholds.

3

Time-Limited Session

Access granted with live countdown. All operations logged to audit trail.

4

Auto-Revoke

Access automatically expires. No cleanup needed, no lingering permissions.

Access Levels

View Rows Browse data in portal UI
Read-Only Queries SELECT on specified schemas
Apply Migrations Higher bar, requires 2+ approvers

JIT Access Grant

Request ID: jit-2026-01-06-sql-7d3f

Active
Time Remaining

01:22:58

Duration

2 hours

Database

orders-db (prod)

Schema Scope

public, analytics

Mode
read-only
Requestor

bob@acme.com

Audit Trail

Approved by @dba-team (2 of 2 required)

Reason: "Investigate slow checkout queries for incident INC-2841"

Ticket: INC-2841

Auto-revokes in 1h 22m 58s
bob
temporary • read-only • 2h
orders-db
AI Suggestion

Based on recent incidents, the Agent suggests adding analytics schema read access for the on-call rotation group.

A workbench that feels designed

Query console, schema browser, migrations, branches, backups, and more. Every tool you need, thoughtfully arranged.

query.sql orders-db
SELECT
o.id,
o.customer_id,
o.total_amount,
o.status
FROM orders o
WHERE o.status = 'pending'
ORDER BY o.created_at DESC
LIMIT 100;
Tables 12 tables
orders 1.2M rows
users 45K rows
products 8.5K rows
order_items 3.8M rows

Columns in orders

PK id
FK customer_id
total_amount
status
created_at
updated_at

001_create_users_table

Applied 3 days ago

002_create_orders_table

Applied 3 days ago

003_add_status_index

Pending • Requires approval

orders-db

Primary • Production

Active

pr-1842-checkout-flow

Branch • Created 2h ago

perf-testing

Branch • Created 1d ago

Continuous Backups

Point-in-time recovery enabled

Healthy

Now

Restore to any point

Daily Snapshot

Today, 3:00 AM

Daily Snapshot

Yesterday, 3:00 AM

Retention Policy

PITR: 7 days Snapshots: 30 days

Ship with branchable SQL
—without shipping secrets.

Postgres & MySQL compatible databases with app-scoped auto-access, instant branching for preview environments, and AI-powered optimization. Start in minutes.

No credential gymnastics
Instant DB branches
JIT access with audit
AI schema optimization

Trusted by engineering teams at

Company
Company
Company
Company