Independently operated. Not affiliated with Datadog, New Relic, Grafana Labs, Dynatrace, Splunk, or Elastic. Pricing sourced from public pages and may not reflect current rates. Verify on each vendor's pricing page before purchasing.
MonitoringCost.com
Run Calculator

Vendor

Middleware pricing 2026: one blended rate per GB

Verified July 2026

Middleware is a full-stack observability platform (metrics, logs, traces, APM, RUM, synthetics) that prices everything on a single consumption meter: $0.30 per GB, blended across signal types, with no per-host fees. Here is how the meter works, what real workloads cost, and where the model does and does not fit.

TL;DR

Pay As You Go: $0.30/GB blended for metrics, logs, and traces, default 30-day retention. RUM: $1 per 1,000 sessions. Synthetic checks: $1 per 5,000; browser test runs $10 per 1,000. 14-day free trial with unlimited ingestion. BYOC is custom. No per-host fees anywhere on the published plan: the bill tracks data volume, not fleet size.

The pricing model

Gigabytes, not hosts

Most incumbent observability pricing splits into per-host platform fees plus separate meters for each product. Middleware collapses all of it into consumption: one blended per-GB rate for telemetry, plus session and check counts for the front-end products.

Middleware, founded in 2022 and backed by Y Combinator, sells the full observability surface: infrastructure metrics, log management, distributed tracing, APM, Real User Monitoring, synthetic monitoring, and an AI incident agent called OpsAI. The published plan is Pay As You Go. Telemetry (metrics, logs, and traces together) bills at $0.30 per GB ingested, with a default 30-day retention. There is no platform fee, no per-host charge, and no separate indexing meter on top of ingest.

The blend matters more than it first appears. Vendors that price logs, metrics, and traces on separate meters force you to forecast three volumes and manage three overage risks. A single blended rate means the only telemetry number you need to know is total GB per month, and shifting instrumentation between signal types (say, moving from chatty logs to structured traces) does not change the rate you pay.

The front-end products stay off the GB meter. RUM bills at $1 per 1,000 sessions, synthetic checks at $1 per 5,000, and full browser test runs at $10 per 1,000. OpsAI splits its billing: error detection is included, while root cause analysis and automated fixes are charged on tokens used, with no fixed per-token rate published. A 14-day free trial covers all features with unlimited ingestion, and a BYOC tier keeps telemetry in your own cloud account at custom pricing. Verify every figure on the Middleware pricing page before relying on it.

The rate card

Every published meter

Verified against middleware.io/pricing on 23 July 2026.
MeterList rateNotes
Metrics, logs, traces (blended)$0.30/GBOne rate across all three signal types. Default 30-day retention on Pay As You Go.
Real User Monitoring$1 per 1,000 sessionsSession-based, separate from the GB meter.
Synthetic checks$1 per 5,000 checksAPI-style synthetic checks.
Browser test runs$10 per 1,000 runsFull browser-based test executions, priced separately from synthetic checks.
OpsAI error detectionIncludedError detection by the AI agent carries no separate charge.
OpsAI root cause + automated fixesPer tokenCharged on tokens used. Middleware does not publish a fixed per-token rate; it provides an estimate calculator instead.
Free trial$0 for 14 daysAll features, unlimited data ingestion, 14-day retention.
BYOC (bring your own cloud)CustomTelemetry stays in your cloud account. Custom pricing and custom retention; quote required.

Source: middleware.io/pricing, verified 23 July 2026. No per-host, per-agent, or platform base fees appear on the published plan.

Three scenarios

What the meter produces at real scales

List-rate arithmetic at three workload shapes. Every line item is derived from the verified rates above; the working is shown in each line.

Startup, ~200 GB/mo telemetry

  • Metrics + logs + traces (200 GB x $0.30)$60
  • RUM (20K sessions x $1/1K)$20
  • Synthetics (50K checks x $1/5K)$10

Total

~$90/month

A small production stack emitting roughly 7 GB/day across all signals. No per-host fee means adding servers costs nothing until they add data.

Mid-market, ~50 GB/day (1,500 GB/mo)

  • Metrics + logs + traces (1,500 GB x $0.30)$450
  • RUM (500K sessions x $1/1K)$500
  • Synthetics (500K checks x $1/5K)$100
  • Browser tests (5K runs x $10/1K)$50

Total

~$1,100/month

At this scale RUM can rival the entire telemetry line. Session volume, not host count, is the number to forecast.

Scale-up, ~200 GB/day (6,000 GB/mo)

  • Metrics + logs + traces (6,000 GB x $0.30)$1,800
  • RUM (2M sessions x $1/1K)$2,000
  • Synthetics (1M checks x $1/5K)$200

Total

~$4,000/month

List-rate arithmetic only. At this volume the BYOC option and a negotiated commitment are the conversation to have; list Pay As You Go is the ceiling, not the floor.

These totals exclude OpsAI root-cause and auto-fix usage, which is token-billed at rates Middleware does not publish. Treat that line as variable until you have real usage data.

Structural comparison

Per-GB blended vs the incumbent meters

The interesting comparison is not the headline number but the shape of the meter. Incumbent figures below are the list rates already documented on this site's Datadog pricing and Honeycomb pricing pages.

Against a per-host model: Datadog lists infrastructure at $15 to $23 per host per month plus $35 per host for APM (see our Datadog breakdown). Under that model a 50-host fleet pays roughly $2,650 per month for infra and APM before a single log is shipped. Under Middleware's meter the same fleet pays $0 for existing and is billed only on the GB it emits. The inversion cuts both ways: a large fleet of quiet hosts favours per-GB, while a small fleet emitting heavy telemetry can price out worse per-GB than per-host.

Against a split log meter: Datadog charges $0.10/GB to ingest logs and then $1.70 per million events to index them, and the indexing line routinely dwarfs ingest (the 100 GB/day scenario on our Datadog page shows $300 of ingest against $20,400 of indexing). Middleware's $0.30/GB is three times Datadog's ingest rate, but it is the whole log bill: there is no separate indexing meter. Comparing $0.30 against $0.10 alone misreads both price lists.

On the front-end products, the published rates sit below Datadog's list: RUM at $1 per 1,000 sessions against Datadog's $1.50 per 1,000, and synthetic API checks at $1 per 5,000 against $5 per 10,000 runs. Honeycomb, for contrast, meters events rather than GB or sessions, which suits trace-heavy teams with disciplined sampling; the models are different enough that only a workload-specific calculation settles it. Our calculator covers the incumbent vendors side by side.

Where it bites

Three ways a Middleware bill runs ahead of forecast

Unbounded log volume

With no per-host anchor, the bill is a direct function of GB shipped. A debug-level logger left on in production, or a noisy sidecar, converts straight into invoice at $0.30/GB. Volume discipline replaces host counting as the budgeting skill.

RUM session growth

RUM bills per session, so a traffic spike or a marketing win raises the monitoring bill in lockstep. At $1 per 1,000 sessions, two million monthly sessions is a $2,000 line item that has nothing to do with server count.

Token-billed OpsAI usage

Root cause analysis and automated fixes charge on tokens used, and no fixed per-token rate is published. A busy incident month means more agent invocations. Watch this line for the first few billing cycles before assuming it is small.

Fit

Who the model suits, and who it does not

Good fit

Teams consolidating several tools onto one bill with one meter to forecast. Fleets with many hosts but moderate telemetry, where per-host pricing overcharges for existence. Startups that want full-stack coverage (APM, RUM, synthetics included) without per-product SKU sprawl, and teams that can use the 14-day unlimited-ingestion trial as a free sizing exercise.

Weaker fit

Small fleets emitting very heavy telemetry, where a per-host price can undercut $0.30/GB. Teams that need published enterprise rate cards and negotiated-discount transparency before engaging sales, since everything beyond Pay As You Go is custom. Buyers who need every line item fixed in advance: the token-billed OpsAI component is variable by design.

Verify before you buy

Middleware publishes its rates at middleware.io/pricing. The figures above ($0.30/GB blended telemetry, $1 per 1,000 RUM sessions, $1 per 5,000 synthetic checks, $10 per 1,000 browser test runs, 14-day free trial, 30-day default retention) were verified on 23 July 2026. BYOC pricing and OpsAI per-token rates are not published; obtain a quote. See our methodology for how this site researches pricing.

Cross-references

Related pages

/datadog-pricing

Datadog pricing breakdown

/new-relic-pricing

New Relic pricing breakdown

/honeycomb-pricing

Honeycomb pricing breakdown

/apm-pricing-comparison

APM pricing across vendors

/rum-pricing-comparison

RUM pricing across vendors

/comparison

Six-vendor comparison

/calculator

Multi-vendor cost calculator

/reduce-monitoring-costs

Twelve cost-reduction strategies

/methodology

How we research pricing

Frequently asked

How much does Middleware cost?
Middleware's Pay As You Go plan charges a blended $0.30 per GB for metrics, logs, and traces, with a default 30-day retention. Real User Monitoring is $1 per 1,000 sessions, synthetic checks are $1 per 5,000, and browser test runs are $10 per 1,000. There is a 14-day free trial with unlimited ingestion, and a BYOC tier with custom pricing. Verify current rates on middleware.io/pricing before purchasing.
Does Middleware charge per host?
No. There is no per-host or per-agent fee on the published Pay As You Go plan. The bill tracks data volume in GB plus session and check counts for RUM and synthetics. A large fleet of quiet servers costs less than a small fleet of chatty ones, which is the opposite of how per-host vendors like Datadog meter infrastructure monitoring.
What does the Middleware free trial include?
The trial runs 14 days at $0 with access to all features, unlimited data ingestion, and 14-day data retention. After the trial the account moves to Pay As You Go rates. Because ingestion is unlimited during the trial, it doubles as a sizing exercise: two weeks of real telemetry tells you what the monthly GB figure will be.
How is OpsAI billed?
OpsAI is Middleware's AI agent for incident work. Error detection is included at no separate charge. Root cause analysis and automated fixes are charged on tokens used. Middleware does not publish a fixed per-token rate on its pricing page; it offers a calculator for estimates. Budget this line as variable until you have a month of real usage.
What is Middleware's BYOC option?
BYOC (bring your own cloud) runs the data plane in your own cloud account, so telemetry does not leave your environment. Pricing and retention are custom and require a quote. It is aimed at teams with data-residency or egress-cost constraints, and at volumes where Pay As You Go list rates stop making sense.