Slowpoke vs Scout Monitoring (Scout APM)

Scout built its reputation on catching N+1 queries and pointing at the line of code, which is genuinely the right instinct and something most APMs still do badly. Slowpoke does the same thing, then runs EXPLAIN, puts a price on it, and refuses to close the ticket until the number stays down.

The short answer

Which one you should pick

Scout is a good product with a clear point of view, and the closest hosted tool to us in spirit. The differences are hosting, price shape, and what happens after the N+1 is found.

Choose Scout Monitoring if

You want hosted, developer-focused APM with good N+1 detection

  • You are on Rails, Django, Flask, FastAPI, Laravel or Elixir and want maintained, opinionated instrumentation.
  • You want automatic N+1 detection in ActiveRecord, Django ORM, Eloquent or Ecto, backtrace included.
  • You want memory bloat detection, which tells you which actions and jobs grow memory — Slowpoke does not do this.
  • You want errors and logs in the same product, and you would rather not host anything.

Choose Slowpoke if

You want the same instinct, free, and taken further

  • You want the N+1 and the plan, the index to add, and what the whole thing costs per day.
  • You want the database server, the PHP-FPM pools, Redis and the machines in the same list.
  • You do not want log storage to become the thing you are actually paying for.
  • Your data has to stay on your machines.
  • You want the finding to become a mission that closes when the measurement says so.

Side by side

Slowpoke and Scout Monitoring, line by line

Written to be fair, including where Scout Monitoring is the better tool. If a row here is wrong or out of date, that is a bug: open an issue and it gets corrected.

Aspect SlowpokeScout Monitoring
Price model€0, self-hostedTransaction tiers; log storage is the main cost driver
Entry price€0From about $19 a month on build-your-own tiers; named plans from about $161 (list, 2026)
Log storageNot a product we sell1 GB included, about $40 a month per additional 30 GB
Traffic surgesNo meter to overrunAbsorbed rather than billed as overage
Where your data livesYour machines, alwaysScout's cloud
N+1 detectionYes, with the route, the count per request and the lineYes, across ActiveRecord, Django ORM, Eloquent and Ecto
Query plans with a fixReal EXPLAIN, twelve named diagnoses
Priced by business costSeconds of waiting per dayTime consumed, per endpoint
Memory bloat detectionNoYes, per action and job
The machine and the database serverCPU, iowait, swap, OOM, disk, Redis, PHP-FPM poolsApplication, plus basic host metrics
Maps what you inheritedOne command reads each machine
Report for non-engineersEffort, return per hour, budget, DORA
Work that verifies itselfMissions closed by measurement
Errors and logsNot its jobIncluded

Finding the N+1 is the easy half

Scout deserves the reputation. Automatic N+1 detection across four ORMs, with the backtrace that shows which line ran the query in a loop, is more useful than a wall of percentiles and it is something plenty of larger APMs still do worse.

Slowpoke detects the same thing — the route that triggers it, how many executions per request, and the file and line — and then keeps going, because in practice the N+1 is rarely the whole story. It runs a real EXPLAIN next to the database, in a read-only transaction, on statements it has already seen there, and reports in plain words: a full table scan on orders reading 1,204,338 rows to return 20, with the composite index that makes it 24. It also catches the eleven other things that are not N+1 at all — a varchar column compared with an integer so the index is silently unusable, an OFFSET 48,000, a LIKE '%…'.

Watch which meter is actually running

Scout's headline is transaction tiers, and it behaves well there — a traffic surge is absorbed rather than billed as overage, which is a genuinely decent policy. The number to look at before you commit is the log line: public pricing includes 1 GB of logs, and additional storage runs around $40 a month per 30 GB. For a chatty application that is the part of the bill that grows, and it grows for a reason that has nothing to do with performance.

Slowpoke does not sell log storage, does not want your logs, and keeps only what it needs to price things. Retention is whatever your own disk allows.

And then the ticket has to close

Both tools can put a finding in front of an engineer. Neither a backtrace nor a dashboard has ever caused a pull request to exist, and performance tickets are written by people who did not create the debt and closed by a tick that proves nothing.

Slowpoke turns the expensive findings into missions with an owner, an effort estimate and a finish line, and will not mark one done until the measured cost has stayed down for three windows — whether that came from the fix, an index someone added, or simply a quiet week, and the history says which. A deploy that doubles a cost becomes a boss linked to that commit. And the mission is readable over MCP, so the assistant your team already uses can pick it up with the query, the plan, the file and the line already in hand.

Can you run both?

Yes. They overlap more than most pairs on this site, so the honest version is: if Scout is already paid for and your estate is one application, you may not need both. If you have servers, databases and queues outside that application, or you want the priced backlog and the report for management, Slowpoke adds what Scout does not and costs nothing to try.

FAQ

Slowpoke and Scout, asked plainly

Does Slowpoke detect N+1 queries like Scout?
Yes: the route that triggers them, how many executions per request, and the file and line that runs them in a loop. It also runs EXPLAIN and reports eleven other named diagnoses that are not N+1, each with the fix it suggests.
Is Slowpoke a free Scout alternative?
For performance and query work, yes, and it is self-hosted. Scout also includes errors, logs and memory bloat detection, none of which Slowpoke does.
Does Slowpoke work with Rails?
Through access logs, database statistics and OpenTelemetry, yes — endpoints, queries and costs, and the agent recognises a Rails application when it reads a machine. A framework package giving the file and line of each query exists for Laravel today, with Symfony, Django, FastAPI and Flask in progress; Rails is not one of them yet.
What does Slowpoke store, and for how long?
Aggregated costs, fingerprinted statements with redacted samples, plans and per-day figures, on your own SQLite or PostgreSQL. Retention is whatever your disk allows, and per-minute metrics are kept for two weeks.

Find out what you inherited

Fifteen minutes on one server, and the first number is yours — it costs nothing to find out.