Maple vs Grafana

Grafana is the dashboard most engineers already know, and behind it sits a stack: Loki for logs, Tempo for traces, Mimir or Prometheus for metrics, each with its own query language. Maple is a source-available observability platform that stores all three signals in one place behind one query surface.

Maple
maple.dev
Grafana
grafana.com

Surfaces

What Grafana leaves you to build

Grafana is a front end over Loki, Tempo and Mimir, each with its own query language. Maple keeps every signal in one store, keyed by the same trace id, with these views already built.

Traces

Every span kept with its attributes and timing, so you see exactly what made a request slow. Click a span for the SQL it ran and the status it returned.

On Grafana: Tempo, deployed, scaled and upgraded on its own, and queried in TraceQL.

See the surface →
POST /checkout 8 spans · 203ms
DurationStatus
/checkoutapi-gateway
203ms200
validateTokenauth-service
24.0msOk
createOrderorder-service
111msOk
validateInventory
20.0msOk
reserveItems
16.0msOk
processPaymentpayment-svc
57.0msOk
chargeCard
45.0msOk
sendConfirmationorder-service
30.0msOk

Logs

Structured lines, linked to the trace that wrote them and kept in the same store as your traces and metrics. Any result opens the request behind it.

On Grafana: Loki, a second backend with its own query language, LogQL.

See the surface →
/logs production
TimeLevelMessage
INFOPOST /api/checkout 402
INFOorder canceled order_id=ord_8421 reason=payment_declined
WARNfalling back to authorize-only flow
ERRORcard declined: insufficient_funds
INFOcharging card amount=129.00 currency=EUR
DEBUGevaluate flag fast_checkout=false cohort=A2
INFOreserved 3 items sku=SKU-9081,SKU-1142,SKU-0077
WARNstock low sku=SKU-9081 qty=3
INFOtoken verified user=u_4109 scope=checkout
DEBUGjwks cache hit kid=2026-05
INFOPOST /api/checkout received
INFOGET /api/cart 200
INFOloaded cart user=u_4109 items=3
INFOGET /api/inventory 200
DEBUGreschedule retry_at=12:04:25 job=sync-inventory
INFOPOST /api/feedback 201
16 logs

Maple on your laptop

One binary with an embedded ClickHouse. `maple start` runs the platform locally with no containers and no account, and self-hosting is one container on your own infrastructure.

On Grafana: self-hosting is mature, with Helm charts, but it means operating four services: Grafana, Loki, Tempo and Mimir.

See the surface →
local · 127.0.0.1:4318 DEMO
maple start
🍁 maple · local mode
listening on http://127.0.0.1:4318
OTLP/HTTP POST /v1/{traces,logs,metrics}
query POST /local/query
dashboard https://local.maple.dev
· bundled UI: pass --offline
data ~/.maple/data
pid 48213 · stop with maple stop
one binary · one port telemetry stays on 127.0.0.1

Differences

Where Maple and Grafana differ

The marker beside each row shows which tool is ahead on it. Rows where both do the same thing are grouped into one line at the end.

Where Maple and Grafana differ
Topic Maple Grafana Edge
Where they differ
Backends to run One system. Traces, logs and metrics land in one store, keyed by the same trace id. Loki, Tempo and Mimir or Prometheus, each deployed, scaled and upgraded on its own, with Grafana in front. Maple
Query languages One visual builder across every signal, SQL through the API and the MCP server when you want it. PromQL for metrics, LogQL for logs, TraceQL for traces. Maple
Pricing model $39 a month with 100 GB per signal included, then $0.30 per GB. No per-user fee. Free tier, then $19 a month plus $8 per active user, $6.50 per thousand active metric series, and $0.55 per GB of logs and traces: $0.05 to process, $0.40 to write and $0.10 to retain for the 30 days. 1 Maple
Per-seat fees Unlimited. Every engineer, on-call responder and manager gets a login. $8 per active user beyond the three included on Pro. 1 Maple
Time to first trace Point an OTLP exporter at Maple and traces start arriving. Provision each backend, connect each as a data source, then build the first dashboard. Maple
Source code On GitHub under FSL-1.1, converting to Apache 2.0 two years after each release. One repository for the whole platform. AGPL-3.0 for Grafana, Loki, Tempo and Mimir. Grafana Cloud itself is proprietary. 2 Even
Self-hosting One deployment: the platform and its database. Self-hostable with mature Helm charts, at the cost of operating four services. Even
OpenTelemetry OTLP in. Nothing else to translate. OTLP through Alloy or the Collector, fanned out to Tempo, Loki and Mimir. 3 Even
Where Grafana is ahead
Plugins No plugin system. Dashboards are built from Maple's own panels. Hundreds of panel and data-source plugins, dashboards-as-code, and a community library of dashboards for almost anything. 4 Grafana
PromQL compatibility PromQL for metrics on Maple Cloud. Grafana dashboards import with each panel's first query; Prometheus alert rules don't import yet. PromQL is the standard for metrics queries. Existing dashboards and recording rules carry over unchanged. Grafana
Non-telemetry data sources Telemetry only. Query Postgres, Elasticsearch, CloudWatch and dozens more from one panel. 4 Grafana
On-call On-call schedules with overrides, escalation chains that page until someone acknowledges, and maintenance windows, with no per-responder fee. Pages go out by email and Slack or Discord message, not phone or SMS. Grafana IRM with on-call schedules, escalation chains and incident timelines. 5 Grafana
Both have
  • Distributed tracing
  • Log management
  • Metrics and dashboards
  • Alerting
  • Error tracking
  • Kubernetes monitoring
  • Public API
  • MCP server for agents

In production : Superwall sends Maple more than 80 billion traces, logs and metrics a month, about 18 TB, with nothing sampled out. Read the Superwall story

Enterprise : Custom volume pricing and retention, priority support, SSO (SAML/OIDC) with SCIM provisioning, and Bring Your Own Cloud to run Maple in your own cloud account. See pricing

Surfaces

Service map

Every service, queue and database, drawn from the spans themselves. There is no topology file to maintain. Edges carry throughput, error rate and latency, and clicking one opens the traces behind it.

See the surface
The Maple service map: services, their dependencies and the request flow between them

Price

The same month, billed twice

A reference workload priced on both at list price. Change the sliders underneath for your own numbers.

The Maple overview page: the logs, traces and metrics received, over request volume and latency
What the meter reads: logs, traces and metrics received, on Maple's overview page.

Reference workload Active metric series 50 k series · Log volume 100 GB/mo · Trace volume 100 GB/mo · Team size 10 users

Maple

$74 per month

Startup plan 100 GB per signal included $39
Overage 116 GB over × $0.30 $35
Team seats No per-seat fees Free
Total $74

Grafana Cloud

$390 per month

Platform fee Base plan $19
Metrics 50k series (10k free) $260
Logs 100 GB (50 GB free) $28
Traces 100 GB (50 GB free) $28
Users 10 users × $8 (3 free) $56
Total $390

$316 less per month on Maple · $3.8k a year

Adjust for your usage

50 k series
100 GB/mo
100 GB/mo
10 users
Maple
$74/mo
Grafana Cloud
$390/mo
Difference
$316 less per month on Maple$3.8k a year · 81% less

List prices checked September 2026. Sources at the foot of the page. Maple pricing based on the Startup plan ($39/mo with 100 GB included per signal (logs, traces, metrics), then $0.30/GB, billed per signal), metered on uncompressed (decoded OTLP) bytes. Grafana bills active series at one data point a minute, so the Maple estimate converts 1k series to about 4.32 GB a month at 0.1 KB per decoded data point. Grafana's $0.55/GB is process, write and retain at the plan's 30 days.

Migration

Moving from Grafana

Telemetry moves with one exporter change. Dashboards and alert rules don't cross vendors, so plan those as the real work.

  1. 01

    Re-point the Collector

    If Alloy or the OpenTelemetry Collector already feeds Tempo, Loki and Mimir, add an OTLP exporter to Maple and run both until the alerts you care about fire from Maple.

  2. 02

    Import the dashboards that matter

    Grafana JSON imports: each panel keeps its first query, rows become sections, and an import report lists anything Maple cannot run yet, extra queries included. Read log panels' lines on the Logs page, rebuild the rest in Maple's builder, and let the long tail of unused dashboards lapse.

  3. 03

    Decommission the stack

    Turn off Loki, Tempo, Mimir and Grafana once Maple is the source of truth. That is four services fewer to upgrade, scale and page on.

otel-collector.yaml The change
exporters:  otlp/tempo:    endpoint: tempo:4317  otlp_http/maple:    endpoint: https://ingest.maple.dev    headers:      x-maple-ingest-key: ${env:MAPLE_INGEST_KEY} service:  pipelines:    traces:      exporters: [otlp/tempo, otlp_http/maple]    logs:      exporters: [otlp/tempo, otlp_http/maple]    metrics:      exporters: [otlp/tempo, otlp_http/maple]
Add Maple as a second exporter, run both, then drop the first.
FAQ

Common questions

Does Maple replace the whole Grafana stack?
Yes. Maple replaces Grafana for dashboards and alerting, Loki for logs, Tempo for traces and Mimir or Prometheus for metrics with one platform. What it doesn't replace is Grafana as a general-purpose charting tool for non-telemetry data sources.
How is Maple different from Grafana Cloud?
Grafana Cloud operates Loki, Tempo and Mimir for you, but you still query three stores in three languages and pay per active user. Maple stores every signal in one place, correlates them by trace id, and has no per-user line.
Do I need to learn a new query language?
No. Maple's query builder works the same way for traces, logs and metrics, and SQL is available through the API and the MCP server. For metrics Maple also runs PromQL, so existing dashboards keep their queries. There is no LogQL or TraceQL to learn.
Can I still self-host?
Yes. Maple is on GitHub under FSL-1.1, converting to Apache 2.0. The difference from the LGTM stack is that you deploy one platform and one database instead of four services with their own scaling models. The self-hosted version has no dashboards, PromQL or alerting yet; those run on Maple Cloud.
What about Grafana's plugins?
Maple has no plugin ecosystem. If your dashboards lean on community panels or on querying business databases alongside telemetry, Grafana is ahead here. Most observability dashboards use a dozen panel types, and Maple's builder covers those.
The Maple services page: latency percentiles, error rate and throughput per service
The services page: latency percentiles, error rate and throughput for every service.

Sources

Where the Grafana claims come from

Vendor facts on this page link to the vendor's own pages and carry the month they were last checked. If one is stale, tell us.

  1. 1 Grafana Cloud pricing checked September 2026
  2. 2 Grafana licensing (AGPL-3.0) checked August 2026
  3. 3 Grafana Alloy checked August 2026
  4. 4 Grafana plugins catalog checked August 2026
  5. 5 Grafana IRM checked August 2026

Compare

Other comparisons

Engraving of a long Roman aqueduct crossing a valley

See the difference on your own traces.

Point one exporter at Maple and run both for a week. The free trial needs a card, with no charge until it ends.