Slowpoke vs New Relic

New Relic charges two meters at once: about $0.40 per gigabyte ingested, and a licence for each person who logs in. Slowpoke has neither. It runs on your machines, the whole team can open it, and the only thing it measures is how much time your users lose.

The short answer

Which one you should pick

New Relic is a broad, mature platform and its free tier is genuinely generous. The question is what happens when the team grows and the data does.

Choose New Relic if

You need one platform and can pay per seat

  • You want APM, infrastructure, logs, browser and mobile monitoring under one login.
  • You need alerting, incident workflows and SLO tracking as part of the same product.
  • Your team is small enough, or your ingest low enough, that the free tier actually covers you.
  • You would rather buy maturity than run anything yourself.

Choose Slowpoke if

You would rather not pay per person who looks

  • You want every engineer to be able to see what is slow, without a licence decision each time.
  • Your ingest is the thing you are trying to reduce, and you do not want it metered.
  • You need the infrastructure you inherited mapped for you, not an empty dashboard to fill.
  • You want the expensive findings turned into work with an owner and a finish line.
  • Your data stays on your machines, and that is not negotiable.

Side by side

Slowpoke and New Relic, line by line

Written to be fair, including where New Relic 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 If a row here is wrong or out of date, that is a bug: open an issue and it gets corrected.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 SlowpokeNew Relic
What it answersWhat it costs you, and what to fix firstWhat is happening, across a broad platform
Price modelNo meter of any kindAbout $0.40 per GB ingested ($0.60 for Data Plus), plus per-user licences
Cost of more engineers looking€0A Full Platform or Core user licence each
Cost of more traffic€0More gigabytes ingested
Free tierThe whole thing, forever100 GB a month and one full user
Where your data livesYour machines, alwaysNew Relic's cloud
Prioritised for youOne list, in seconds of waiting per dayCharts, alerts and NRQL you write
Query plans with a fixReal EXPLAIN, twelve named diagnosesSlow query traces, no recommendation
Maps what you inheritedOne command reads each machineYou tell it what to instrument
Says what it cannot seeYes, with the lines that fix it
Work that verifies itselfMissions closed by measurement
Live incident responseNot its jobAlerting, SLOs, incident workflows
BreadthPerformance debt onlyAPM, infra, logs, browser, mobile, synthetics

Two meters, and one of them punishes the behaviour you want

New Relic bills on data ingested — roughly $0.40 per gigabyte, or $0.60 on Data Plus — and separately on people. Users are split into Full Platform and Core tiers, and a fifteen-person engineering team with five full and ten core users runs to a four-figure monthly sum in licences alone, before a single gigabyte is counted.

The ingest meter has the same awkward shape as any usage-based observability pricing: the busier and slower your system gets, the more it costs to watch. The seat meter is worse, because it prices the thing you are trying to encourage. A performance culture means more engineers looking at performance data more often, and per-seat licensing makes every one of those people a line item.

Slowpoke has no seats and no ingest meter. It is open source and self-hosted, so the entire team can have an account and the cost does not move.

A platform you fill in, against a tool that arrives with an answer

New Relic gives you an enormous, capable surface and a query language to interrogate it. That is the right product if you already know what you are looking for.

If you have just inherited the system, you do not. Slowpoke starts from the other end: one command per machine reads what is there — nginx sites, PHP-FPM pools, MySQL and PostgreSQL, Redis, and the applications with their framework — writes its own configuration, and draws the map. Fifteen minutes later there is a list, ordered by seconds of waiting per day, and you did not have to know which question to ask.

The part neither platform does: making the fix happen

Finding the slow query has not been the hard part for years. Getting it fixed is. Debt tickets are written by people who did not create the debt, sit in a backlog, and get closed by a tick in Jira that proves nothing.

Slowpoke turns the expensive items into missions with an owner, an effort estimate and a finish line, and refuses to call one done until the measured cost has stayed down for three windows. A deploy that makes something twice as slow becomes a boss with that commit's name on it. This is the part that makes the numbers turn into shipped changes, and no APM does it.

Can you run both?

Yes. Slowpoke never sits in the request path and reads what your servers already write, so it adds nothing to your ingest bill and cannot conflict with a New Relic agent. Several teams use the Slowpoke debt report precisely to argue for the budget the New Relic contract sits in.

FAQ

Slowpoke and New Relic, asked plainly

Is Slowpoke a free New Relic alternative?
For finding and prioritising slow endpoints and slow queries, yes, and it is self-hosted so nothing is metered. It does not replace New Relic's alerting, logs, browser and mobile monitoring, or incident workflows.
Does Slowpoke charge per user?
No. There is no seat count, no per-host price and no ingest meter. Everyone on the team can have an account on the instance you run.
Will Slowpoke increase my New Relic bill?
No. Slowpoke reads access logs, slow logs, database statistics and /proc on your own machines and sends to your own server. It does not emit telemetry to any vendor.
Can I keep New Relic and still use Slowpoke?
Yes, and it is a normal setup. The agent is read-only, capped at one CPU core and 64 MB, and never sits in the request path, so it does not interfere with another vendor's agent.

Find out what you inherited

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