Introducing Supabase Observability Triggers

Introducing Supabase Observability Triggers

Introducing Supabase Observability Triggers

Justin Crites

Justin Crites

Product

5

min read


We’ve just launched Supabase Observability Triggers. A trigger watches any stream of data that comes through Dreambase, checks it against a threshold you set, and sends the result to the destination you choose: your inbox, a Slack channel, a webhook, or an AI agent. This is alerts for the agentic era. Here's a few ideas of what you can do with them:

  • Research and resolve incidents before they grow: Send your agent a trigger when CPU sits at 80% for 10 minutes, and a critical alert when the disk passes 95%. It'll automatically start researching what happened and why using the context you've already curated in it. If you have the Supabase MCP enabled it can take actions so that your users never have downtime.

  • React when the business changes in realtime: Set up a metric to watch critical parts of your business and get notified faster than looking at a dashboard. You ship a new release Thursday, signups drop off 20% within the first hour, but you get a notification in slack right after your threshold. Now you can patch up the bug that caused it with little downtime.

  • Build custom observability for a release. Shipping a new feature this week, and worried about storage costs? Set up a trigger to watch your disk utilization for this release and ping the agent who built the feature if things are getting out of control.

Watch any data source in Dreambase


Triggers are built to watch any stream of data. Internally we have a few use cases you can hook up now:

Live database metrics: For each connected Supabase project you can pull in CPU, memory, disk, connections in use, WAL size, database size, transaction ID age, Data API pool timeouts, and how long the database has gone without answering.

Health checks: Each time Dreambase checks a project for estimated monthly bill, plan and compute price, monthly active users, total users, storage size, open and priority recommendations, and the health score you can build actions right off of them.

Metrics: Any metric in your workspace can be built into a trigger. Signups, revenue, active teams, social listening data, or anything that you care about can be watched on the same schedule as the underlying datasets refresh.

One trigger can read also across all of these! A CPU trigger can also check connections, a rule on one project can compare it with another, or A database signal can be measured against a business metric.

We've also made it easy to get started with observability. Every Supabase project you connect comes with 13 default triggers, tuned to Supabase’s own guidance. They’re just triggers, so edit them to your liking, delete them outright, or write custom ones yourself.

Thresholds as simple or as smart as you need


A basic alert is a simple line like: value >= 80. Not here. We've deployed a small formula language built for triggers, and you can take it much further than a single line:

You want to know when…

Formula

CPU stays high, not just spikes

held(value >= 80, "10m")

CPU and connections are both high, on a CPU trigger

held(value > 85 && connectionsPercent > 80, "10m")

More than 5 pool timeouts happen in 20 minutes

sum(value, "20m") > 5

The database grew 20% in 12 hours

change(value, "12h") >= 20

Signups fell 30% compared with a week ago

change(value, "7d") <= -30

The health score dropped 10 points since the last check

delta(value) <= -10

A signal has gone quiet for 30 minutes

silence(value) >= 30

Database size per user crosses a line

value / metric("<users metric id>") > 1000000

You can combine conditions with &&, || and !(…), do arithmetic with + - * /, and read another project’s signals with project("<id>"). Each trigger can have a warning level and a critical level, so an alert escalates as things get worse, and a clear condition, so it doesn’t flap on and off around the line. A gap in the data never opens or closes an alert by accident.

And you can backtest it. Dreambase replays the rule against up to 14 days of minute data, 90 days of health checks or a year of metric history, and tells you how often it would have fired.

Send it anywhere, including to an agent

You choose the destination for each trigger, and you can route by severity: warnings to Slack, critical alerts to email.

  • Email to the owners and admins you pick.

  • Slack through an incoming webhook.

  • Signed webhooks to your own endpoint, to start any workflow you already run.

  • Claude routines and Grok Bot. When one of the default Supabase triggers fires, the agent picks up the alert and starts investigating.

  • MCP Events, supported natively. Any agent connected to Dreambase can subscribe to the trigger event and hear every alert it’s allowed to see as it fires, with no destination to set up.

This turns an alert into the first step of a workflow. A webhook can open a ticket, a routine can start a diagnosis, and an agent can read the firing, pull the project’s metrics and start a fix.k

Set it up through MCP

Triggers are built to be set up by an agent.

  1. Connect the Dreambase MCP server to Claude, ChatGPT, Cursor, Codex or any MCP client: add https://app.dreambase.com/mcp, or install the Dreambase plugin.

  2. In Dreambase, open a Supabase project’s Settings tab and click Add next to Custom. This copies a prompt for that project.

  3. Paste it into your assistant and say what you want to watch, in plain English.

The agent uses the same tools you’d call yourself: list_triggers, create_trigger, update_trigger, delete_trigger, test_trigger and list_trigger_firings. It suggests a signal and levels, backtests them, and asks before it saves anything. The same tools are on the REST API, so you can keep your alerting in code too.

Get started

Triggers are live now for workspace owners and admins with Observability. Connect a Supabase project, open its Settings tab, turn on the defaults, then ask your agent for your first custom trigger. Every default, signal and formula is covered in the Triggers docs.

Want triggers on your business metrics? Analytics is set up with our team. Book a call and we’ll get your metrics flowing.

What do you want to know from your data?

Get the answer now.

Ready to unlock
your Supabase data?

Analytics, dashboards, and reports are set up with our team, so meet with us to get started. Want to monitor your Supabase project today? Start for free.

Ready to unlock
your Supabase data?

Analytics, dashboards, and reports are set up with our team, so meet with us to get started. Want to monitor your Supabase project today? Start for free.