Commercial SaaS
Datadog, New Relic, Splunk. Three signals are correlated, but the bill grows linearly. Mid-size teams pay $2k–10k/mo for 100 GB/day. Reduce intake and you lose visibility.
comparison
Why molesignal
The three observability paths, and why we built a fourth.
Three paths, each with a tax
OSS stack
Loki + Mimir + Tempo + Grafana. Free, but logs / metrics / traces live in three stores with three query languages. The trace ↔ log ↔ host metric jump is manual.
molesignal
One storage layer (Parquet on object store), one query engine (DataFusion + Arrow), one metadata layer (Postgres). Three signals correlated at the data plane. Self-hosted.
Side by side
Every claim is testable. Spin up the sandbox, send your own data, and reproduce.
100 GB/day cost
- Commercial SaaS
- ~$2k+/mo · grows linearly
- OSS stack
- infra only
- molesignal
- infra only
Three signals — same storage
Whether logs, metrics, traces share one storage layer.
- Commercial SaaS
- ✓ (their cloud)
- OSS stack
- 3 stores, 3 query langs
- molesignal
- Parquet + DataFusion
Cross-signal correlation
Jump from a trace to its logs to the host metric — without copy/paste.
- Commercial SaaS
- ✓ (paid)
- OSS stack
- manual trace_id copy-paste
- molesignal
- native (/web/correlation/*)
Data ownership
- Commercial SaaS
- their cloud
- OSS stack
- self-hosted
- molesignal
- self-hosted
Setup time
- Commercial SaaS
- 5 min (agents)
- OSS stack
- 6 hours+ (5 components + Grafana)
- molesignal
- 1 cmd `docker compose up`
OpenTelemetry-native
- Commercial SaaS
- yes
- OSS stack
- partial
- molesignal
- yes (9 intake protocols)
Real-time alerts (<1s)
- Commercial SaaS
- yes
- OSS stack
- no (eval ≥ scrape interval)
- molesignal
- yes (kind: realtime)
Multi-tenant out-of-box
- Commercial SaaS
- yes (per-account)
- OSS stack
- no
- molesignal
- yes (planner-level org rewrite)
Run the numbers
Pick your real-world intake volume and retention. We compute Datadog at public pricing and molesignal as S3 + a small compute baseline.
Datadog (est.)
$7.4k
per month
molesignal
$150
per month (S3 + compute)
Savings
$7.2k
98% less
Order-of-magnitude estimates based on Datadog's public list pricing as of June 2026. Datadog: logs intake + 15-day indexed events (assuming all received logs are indexed — an upper bound) + APM spans, with a retention multiplier. molesignal: S3 Standard storage + baseline Postgres + compute scaling per 100 GB/day. Your mileage will vary with redundancy, region, commit discounts, and workload shape. Source: Datadog public pricing.
Five scenarios where this matters
- 01 / 05
3am, page just fired
SRE jumps three tabs to stitch a trace_id across logs, metrics, and traces. molesignal makes that one query.
- 02 / 05
Budget cut by 30%
CFO wants observability spend down. SaaS forces you to reduce intake; molesignal lets you keep visibility, change the cloud bill.
- 03 / 05
Stuck maintaining ES/CK clusters
Self-built scales operationally as fast as it scales data. molesignal is one set of components on Parquet + object storage.
- 04 / 05
Multiple teams, one platform
Planner-level org-rewrite means you don't deploy per-team. One process, many tenants, AI-grade isolation.
- 05 / 05
Alert latency matters
Prometheus eval-interval ≥ scrape-interval. With kind: realtime alerts, molesignal pages under a second.
What molesignal is not
We're an observability data plane — not a hosted SaaS in v1 and not magic. You still choose how applications are instrumented and how the platform is operated.