What is a COLLSCAN and why should I care?

COLLSCAN means MongoDB scanned every document in a collection to satisfy a query because no usable index was found. On busy collections this is one of the most common causes of slow Meteor apps. SkySignal flags every COLLSCAN so you can add the missing index.

Does SkySignal show queries inside publications?

Yes. Queries fired from publications, methods and background jobs are all captured and attributed to the originating source so you know which code path is causing the load.

Can Astra suggest indexes?

Yes. Astra AI analyzes query shapes and proposes compound indexes with customerId first for multi-tenant apps. It can open a pull request with the migration.

Feature - Database Insights

MongoDB Query Insights for Meteor

Every slow query, every COLLSCAN, every aggregation that stalled a method - captured in context so you know exactly which publication or handler to fix.

Start Free Trial What we track

Mongo is where Meteor apps live

Query visibility scoped to the code that called it

A database monitor that only shows you the top N queries is only useful once you figure out which application code fired them. In Meteor that is not obvious - the same findOne can run from a method, a publication, or a Steve Jobs background task. SkySignal captures the originating caller with every sample so the database tab answers "who" as well as "what".

Because the agent traces from inside your Meteor process, it sees exactly what your code asked for, not what Mongo happened to log. Multi-tenant apps see every query filtered by tenant, and admins get a global roll-up.

Every query signal that matters

Slow query detection

Every query crossing your configured threshold is captured with collection, filter shape, duration and the method or publication that issued it.

COLLSCAN flagging

Queries that scanned every document in a collection get a loud red flag so you can add the missing index before traffic grows.

Slow aggregation pipelines

Aggregations are captured with their full pipeline so you can see the stage that is dragging, tune it and ship.

Write frequency per collection

See which collections are being written to most, surface hot documents and catch loops that accidentally hammer Mongo.

COLLSCAN detection

The single most useful query flag

When MongoDB runs a query without an index, it scans the whole collection. On a 10k document collection that is a shrug. On a 10M document collection it is a page that never loads. SkySignal checks every slow query for the COLLSCANplan stage and shows it next to the source file and line that triggered it, so the fix is always one pull request away.

COLLSCAN detected

// orders.js  method: orders.list
Orders.find({
  customerId,
  status: 'open'      // <- no index on status
}).fetch()

plan: COLLSCAN
docs examined: 1,284,331
duration: 4,812ms

Astra can open a PR adding { customerId: 1, status: 1 } with migration notes.

How it works

In-process driver instrumentation

Wrap the driver

The agent instruments the Mongo driver so every find, update, aggregate and command is timed and fingerprinted.

Attribute to caller

Trace context propagates through AsyncLocalStorage, so each sample knows whether a method, publication or job fired it.

Plan and rank

Slow queries are explained to detect COLLSCAN and ranked by total time contribution so you fix what matters most.

Connection pool and topology

See active and idle pool size, checkout wait times and topology changes. Spot pool exhaustion before it cascades into method timeouts.

AI-suggested indexes

Astra reads the flagged queries, proposes compound indexes with multi-tenant customerId first and opens a PR with the migration script for review.

Make MongoDB fast again

Install the agent, open the Database tab, fix the red rows.

Start Free Trial Method tracing