Observability

Projects

Projects

One Supabase project's health, recommendations, usage, alerts and settings.

The Projects page shows all Dreambase data for a single Supabase project: health score, recommendations, cost, alert history, and alert destinations.

Location: Observability > Projects. The page opens on the linked project, otherwise the last project viewed, otherwise the project that most needs attention. The picker at the top switches projects. Available to workspace owners and admins.

The tabs

Overview

Health score and grade, open recommendations by category, recent activity such as deploys, migrations, and resizes, and bill and capacity. Two setup cards link to Set up Alerts and Automate with Agents.

Project Overview tab with health grade and open recommendations

Recommendations

Every finding from the project's checks, grouped by category with each category's score. Selected findings can be acted on from the Take Action toolbar:

  • Fix in Supabase: opens the relevant Supabase dashboard page.

  • Migration: copies reviewed fix SQL as a migration.

  • Pull Request: sends the fix to the project's repository.

  • Validate: re-checks the finding immediately.

  • Won't fix: dismisses the finding and records the reason.

  • Inspect with agent: hands the finding to an agent with full context.

Upgrades lists recommended plan or compute changes. History shows every finding's past states.

Project Recommendations tab with findings grouped by category

Usage & Cost

Resource usage and cost for compute, disk, egress, and other line items, measured against the plan's limits.

Project Usage & Cost tab with compute, disk and egress cost

Alerts

Every alert raised by the project in the last 30 days, with start and clear times.

Settings

Alert routing and thresholds. Notification destinations (email, Slack, webhook, Claude routine, Grok Bot) are added here, and the project’s triggers are turned on and tuned. Agents can also subscribe through MCP Events. See Triggers.

Project Settings tab with alert routing and destinations

How checks work

Every project receives an initial check on connection. After that, an hourly check covers health and performance, and a daily check covers all categories: security, performance, indexing, errors, and capacity. A finding deducts 25 points when critical, 5 for a warning, and 1 for a minor note.

Checks are read-only. A fix runs only when a user applies it.

Tool reference

get_supabase_monitoring

MCP and REST · health-reports:read

{ workspaceId, connectionId, view?, range?, asOf? }

The project’s health at a glance: grade, category scores, open findings and recent checks. view: "detail" adds history over range: 7D, 30D (the default) or 90D. asOf reads the project as it was at an earlier time.

get_supabase_health_history

MCP and REST · health-reports:read

{ workspaceId, connectionId, view?, days? }

The project’s score and grade over time, one point per check.

get_engineering_alerts

MCP and REST · health-reports:read

{ workspaceId, connectionId?, severity?, category?, view?, limit?, cursor? }

Lists open findings across the workspace or for one project. Filter by severity and category.

get_engineering_alert_detail

MCP and REST · health-reports:read

{ workspaceId, connectionId, key }

One finding in full: what was found, why it matters, the affected objects and the fix.

set_engineering_alert_states

MCP and REST · health-reports:write

{ workspaceId, connectionId, keys, state, reason? }

Sets findings to done (fixed, waiting for the next check to confirm), fixing (the fix is under way), dismissed (won’t fix, with a reason) or back to open.

validate_engineering_alert

MCP and REST · health-reports:write

{ workspaceId, connectionId, key }

Re-checks one finding immediately and returns whether it is fixed.

run_supabase_check

REST only · health-reports:write

{ workspaceId, connectionId }

Runs the project’s full check immediately instead of at the next scheduled time.

list_supabase_branches

MCP and REST · health-reports:read

{ workspaceId, connectionId }

Lists the project’s Supabase branches and whether each is running.

Read a project’s health through the API

curl "https://api.dreambase.com/v1/workspaces/<workspace-id>/connections/supabase/<connection-id>/monitoring?range=30D" \
  -H "Authorization: Bearer <access-token>"

REST paths sit under /workspaces/{workspace_id}/connections/supabase/{connection_id}/.

Important boundaries

  • The page and its tools are restricted to owners and admins.

  • Findings reflect the latest check. The hourly check updates health and performance; all other categories update daily.

  • Dreambase makes no changes to the project. A fix runs only when applied by a user, in Supabase or through a migration.

  • A dismissed finding stays dismissed until it is reopened.

The Projects page shows all Dreambase data for a single Supabase project: health score, recommendations, cost, alert history, and alert destinations.

Location: Observability > Projects. The page opens on the linked project, otherwise the last project viewed, otherwise the project that most needs attention. The picker at the top switches projects. Available to workspace owners and admins.

The tabs

Overview

Health score and grade, open recommendations by category, recent activity such as deploys, migrations, and resizes, and bill and capacity. Two setup cards link to Set up Alerts and Automate with Agents.

Project Overview tab with health grade and open recommendations

Recommendations

Every finding from the project's checks, grouped by category with each category's score. Selected findings can be acted on from the Take Action toolbar:

  • Fix in Supabase: opens the relevant Supabase dashboard page.

  • Migration: copies reviewed fix SQL as a migration.

  • Pull Request: sends the fix to the project's repository.

  • Validate: re-checks the finding immediately.

  • Won't fix: dismisses the finding and records the reason.

  • Inspect with agent: hands the finding to an agent with full context.

Upgrades lists recommended plan or compute changes. History shows every finding's past states.

Project Recommendations tab with findings grouped by category

Usage & Cost

Resource usage and cost for compute, disk, egress, and other line items, measured against the plan's limits.

Project Usage & Cost tab with compute, disk and egress cost

Alerts

Every alert raised by the project in the last 30 days, with start and clear times.

Settings

Alert routing and thresholds. Notification destinations (email, Slack, webhook, Claude routine, Grok Bot) are added here, and the project’s triggers are turned on and tuned. Agents can also subscribe through MCP Events. See Triggers.

Project Settings tab with alert routing and destinations

How checks work

Every project receives an initial check on connection. After that, an hourly check covers health and performance, and a daily check covers all categories: security, performance, indexing, errors, and capacity. A finding deducts 25 points when critical, 5 for a warning, and 1 for a minor note.

Checks are read-only. A fix runs only when a user applies it.

Tool reference

get_supabase_monitoring

MCP and REST · health-reports:read

{ workspaceId, connectionId, view?, range?, asOf? }

The project’s health at a glance: grade, category scores, open findings and recent checks. view: "detail" adds history over range: 7D, 30D (the default) or 90D. asOf reads the project as it was at an earlier time.

get_supabase_health_history

MCP and REST · health-reports:read

{ workspaceId, connectionId, view?, days? }

The project’s score and grade over time, one point per check.

get_engineering_alerts

MCP and REST · health-reports:read

{ workspaceId, connectionId?, severity?, category?, view?, limit?, cursor? }

Lists open findings across the workspace or for one project. Filter by severity and category.

get_engineering_alert_detail

MCP and REST · health-reports:read

{ workspaceId, connectionId, key }

One finding in full: what was found, why it matters, the affected objects and the fix.

set_engineering_alert_states

MCP and REST · health-reports:write

{ workspaceId, connectionId, keys, state, reason? }

Sets findings to done (fixed, waiting for the next check to confirm), fixing (the fix is under way), dismissed (won’t fix, with a reason) or back to open.

validate_engineering_alert

MCP and REST · health-reports:write

{ workspaceId, connectionId, key }

Re-checks one finding immediately and returns whether it is fixed.

run_supabase_check

REST only · health-reports:write

{ workspaceId, connectionId }

Runs the project’s full check immediately instead of at the next scheduled time.

list_supabase_branches

MCP and REST · health-reports:read

{ workspaceId, connectionId }

Lists the project’s Supabase branches and whether each is running.

Read a project’s health through the API

curl "https://api.dreambase.com/v1/workspaces/<workspace-id>/connections/supabase/<connection-id>/monitoring?range=30D" \
  -H "Authorization: Bearer <access-token>"

REST paths sit under /workspaces/{workspace_id}/connections/supabase/{connection_id}/.

Important boundaries

  • The page and its tools are restricted to owners and admins.

  • Findings reflect the latest check. The hourly check updates health and performance; all other categories update daily.

  • Dreambase makes no changes to the project. A fix runs only when applied by a user, in Supabase or through a migration.

  • A dismissed finding stays dismissed until it is reopened.

The Projects page shows all Dreambase data for a single Supabase project: health score, recommendations, cost, alert history, and alert destinations.

Location: Observability > Projects. The page opens on the linked project, otherwise the last project viewed, otherwise the project that most needs attention. The picker at the top switches projects. Available to workspace owners and admins.

The tabs

Overview

Health score and grade, open recommendations by category, recent activity such as deploys, migrations, and resizes, and bill and capacity. Two setup cards link to Set up Alerts and Automate with Agents.

Project Overview tab with health grade and open recommendations

Recommendations

Every finding from the project's checks, grouped by category with each category's score. Selected findings can be acted on from the Take Action toolbar:

  • Fix in Supabase: opens the relevant Supabase dashboard page.

  • Migration: copies reviewed fix SQL as a migration.

  • Pull Request: sends the fix to the project's repository.

  • Validate: re-checks the finding immediately.

  • Won't fix: dismisses the finding and records the reason.

  • Inspect with agent: hands the finding to an agent with full context.

Upgrades lists recommended plan or compute changes. History shows every finding's past states.

Project Recommendations tab with findings grouped by category

Usage & Cost

Resource usage and cost for compute, disk, egress, and other line items, measured against the plan's limits.

Project Usage & Cost tab with compute, disk and egress cost

Alerts

Every alert raised by the project in the last 30 days, with start and clear times.

Settings

Alert routing and thresholds. Notification destinations (email, Slack, webhook, Claude routine, Grok Bot) are added here, and the project’s triggers are turned on and tuned. Agents can also subscribe through MCP Events. See Triggers.

Project Settings tab with alert routing and destinations

How checks work

Every project receives an initial check on connection. After that, an hourly check covers health and performance, and a daily check covers all categories: security, performance, indexing, errors, and capacity. A finding deducts 25 points when critical, 5 for a warning, and 1 for a minor note.

Checks are read-only. A fix runs only when a user applies it.

Tool reference

get_supabase_monitoring

MCP and REST · health-reports:read

{ workspaceId, connectionId, view?, range?, asOf? }

The project’s health at a glance: grade, category scores, open findings and recent checks. view: "detail" adds history over range: 7D, 30D (the default) or 90D. asOf reads the project as it was at an earlier time.

get_supabase_health_history

MCP and REST · health-reports:read

{ workspaceId, connectionId, view?, days? }

The project’s score and grade over time, one point per check.

get_engineering_alerts

MCP and REST · health-reports:read

{ workspaceId, connectionId?, severity?, category?, view?, limit?, cursor? }

Lists open findings across the workspace or for one project. Filter by severity and category.

get_engineering_alert_detail

MCP and REST · health-reports:read

{ workspaceId, connectionId, key }

One finding in full: what was found, why it matters, the affected objects and the fix.

set_engineering_alert_states

MCP and REST · health-reports:write

{ workspaceId, connectionId, keys, state, reason? }

Sets findings to done (fixed, waiting for the next check to confirm), fixing (the fix is under way), dismissed (won’t fix, with a reason) or back to open.

validate_engineering_alert

MCP and REST · health-reports:write

{ workspaceId, connectionId, key }

Re-checks one finding immediately and returns whether it is fixed.

run_supabase_check

REST only · health-reports:write

{ workspaceId, connectionId }

Runs the project’s full check immediately instead of at the next scheduled time.

list_supabase_branches

MCP and REST · health-reports:read

{ workspaceId, connectionId }

Lists the project’s Supabase branches and whether each is running.

Read a project’s health through the API

curl "https://api.dreambase.com/v1/workspaces/<workspace-id>/connections/supabase/<connection-id>/monitoring?range=30D" \
  -H "Authorization: Bearer <access-token>"

REST paths sit under /workspaces/{workspace_id}/connections/supabase/{connection_id}/.

Important boundaries

  • The page and its tools are restricted to owners and admins.

  • Findings reflect the latest check. The hourly check updates health and performance; all other categories update daily.

  • Dreambase makes no changes to the project. A fix runs only when applied by a user, in Supabase or through a migration.

  • A dismissed finding stays dismissed until it is reopened.

Organizations

Monitoring

© 2026 Dream, Inc. All rights reserved.