See the machine, not a marketing dial
Platform metrics, Redis state and diagnostics, including the connection accounting that decides whether your next deploy fits.
- Connection budget accounting
- Redis visibility
- Diagnostics on demand
Connections, counted honestly
The telemetry endpoint reports the sum of role limits and how far it overcommits the server. Over-commitment is a warning you can act on, not a silent failure later.
Memory is a hard ceiling
Container memory limits are enforced by the kernel. Monitoring shows you how close a workload runs before the kernel makes the decision for you.
Redis is not a black box
Inspect Redis state from the dashboard rather than guessing from application behaviour.
Diagnostics you can trigger
Run a diagnostic pass when something looks wrong, instead of waiting for a scheduled report that arrives after the incident.
The metric that actually bites
On a single-server platform, connections run out before memory does. PostgreSQL reserves connections for the superuser, every role brings its own limit, and there is no shared ceiling enforcing sanity. Watch the sum, not one pool.
Build the whole backend in one project
One plan, one dashboard, one command line interface. PostgreSQL, applications, functions, storage and realtime.