Blog / Observability
The best Datadog alternatives in 2026, by the reason you’re leaving
Datadog alternatives grouped by why teams switch: unpredictable bills, data residency, open source, deep debugging. Prices from vendors’ own pages.
The right Datadog alternative depends on why you’re leaving. For unpredictable bills, look at tools priced on one simple unit. For data that must stay in your cloud, look at BYOC platforms like Kloudfuse. For open source, look at Grafana or SigNoz. Many teams can switch without re-instrumenting, if they pick a tool that accepts the agents they already run.
Why teams look for Datadog alternatives
Datadog is good software. It’s quick to set up, it has a huge catalog of integrations, and most engineers like using it. Very few people leave because it doesn’t work.
They leave because of the bill. One Capterra reviewer says Datadog’s pricing structure can become “complex and unpredictable as usage grows”. Another, on TrustRadius, points at custom metrics: “Custom metrics rapidly increases the cost and we can’t predict them.”
The reason is structural. Datadog prices each product on its own meter (from Datadog’s pricing page, checked October 2026):
| What you use | How Datadog charges |
|---|---|
| Infrastructure monitoring | $15 (Pro) or $23 (Enterprise) per host per month, billed annually |
| APM | $31 per host per month with Infrastructure, billed annually |
| Log ingestion | $0.10 per GB |
| Log indexing | $1.70 per million events, 15-day retention, billed annually |
| Custom metrics | 100 (Pro) or 200 (Enterprise) included per host; beyond that, billed on usage at a rate from Datadog sales |
Each meter is reasonable on its own. Together, they move independently. A Kubernetes migration adds hosts. A new tag multiplies custom metrics, because each unique tag combination counts as a separate metric. A noisy deploy spikes log volume.
Before you switch: renegotiating is a real option
Switching takes work. Sometimes a better contract is enough.
After a famously large Datadog bill, Coinbase built and tested an in-house Grafana, Prometheus and ClickHouse stack, then renegotiated with Datadog and stayed, as The Pragmatic Engineer reported. If your main problem is price and your team is happy otherwise, take your usage data to Datadog first. Come back to this list if that doesn’t work.
Pick your reason for leaving
“Our bill is unpredictable”
Look for tools that charge on fewer, simpler units, so growth in one area doesn’t multiply costs in another.
- Kloudfuse: priced on one axis, terabytes ingested per day. No per-host, per-user or custom-metric fees. Your total telemetry volume in TB per day sets your price.
- New Relic: $0.40 per GB ingested after 100 GB free each month, plus per-user fees for full platform users. Simple for small teams with few users; user fees add up as teams grow.
- SigNoz: $0.30 per GB for logs and traces and $0.10 per million metric samples on its cloud plan, with a $49 monthly minimum. Open source underneath, built on OpenTelemetry.
“Our data has to stay in our own cloud”
Regulated industries often can’t send telemetry to a third-party SaaS. BYOC (bring your own cloud) platforms run inside your account while the vendor operates them. Kloudfuse isn’t the only option here, and you should compare.
- Kloudfuse: runs as Self-SaaS, our managed BYOC model. The full data plane, including ingestion, storage and queries, runs in your VPC. Only operational metadata such as version and health reaches us. FIPS 140-3, HIPAA and PCI DSS.
- groundcover: BYOC or on-prem only, priced per host (a free plan, and paid plans from $30 to $50 per host per month). Built around eBPF, so it needs little code instrumentation on Kubernetes.
- Last9: single-tenant in your own cloud account, managed by Last9. No public pricing.
- SigNoz Enterprise and Grafana Enterprise also offer BYOC deployments on their enterprise plans.
When comparing, ask where queries run, not just where data is stored. Some offers store data in your bucket but still process it in the vendor’s cloud. Our BYOC buyer’s guide has eight questions to put to any vendor.
“We want open source, or to run it ourselves”
- Grafana: the open-source stack (Grafana, Prometheus or Mimir, Loki, Tempo) can be self-hosted. Grafana Cloud has a free tier, and its Pro plan has a $19 monthly platform fee plus $6.50 per 1,000 active metric series, with logs billed per GB in three parts (process, write, retain) after 50 GB included.
- SigNoz: open source and OpenTelemetry-native, with cloud, self-hosted and BYOC options.
- Elastic Observability: serverless, hosted or self-managed. Serverless log ingestion starts at $0.07 per GB (Logs Essentials) or $0.09 per GB (Complete), with retention billed separately per GB. A natural fit if you already run Elasticsearch.
Self-hosting saves license fees but moves the operating work to your team, which is the trade-off Self-SaaS was designed around. Budget for the people, not just the servers.
“We need to debug high-cardinality data”
- Honeycomb: built for asking new questions of rich event data. Priced on events per month, with Pro from $150 a month.
“We’re a large enterprise and want one suite”
- Dynatrace: strong automatic discovery. Full-stack monitoring is $58 a month per 8 GiB host, and log ingest and processing is $0.20 per GiB, with retention and queries billed separately.
- Splunk Observability: from $15, $60 or $75 per host per month by edition, billed annually. Most natural for teams already on Splunk.
Two market changes to know about: Chronosphere is now part of Palo Alto Networks, and Observe is now part of Snowflake. Check how that affects roadmap and pricing before you commit.
How to compare prices across different units
This is where most comparisons go wrong. One vendor charges per host, another per GB, another per metric series, another per user. You can’t compare the headline numbers directly.
Instead, gather five numbers about your own environment:
- Hosts and containers you monitor
- GB per day of logs, traces and metrics, separately
- Active metric series, including custom tags
- People who need full access
- How long you need to keep data
Then ask each vendor to price that same profile. Price it for next year’s volume as well as today’s, because the tools that look cheapest at small scale aren’t always cheapest when you grow.
If you’re on Kubernetes, pay extra attention to per-host pricing. Containers and autoscaling can change your host count every hour.
How to switch without re-instrumenting
The biggest fear in leaving Datadog is redoing the instrumentation. With some alternatives, you don’t have to.
Keep your agents. Kloudfuse accepts data from the Datadog Agent, the OpenTelemetry Collector and Fluent Bit. To switch a Datadog Agent, you point its endpoints at Kloudfuse instead of Datadog. Our migration guide walks through the settings. One of our customers runs 317 Datadog Agents into Kloudfuse across GCP, AWS and on-prem.
Run both side by side. The Datadog Agent can send to two destinations at once (Kloudfuse needs Agent 7.41 or later). Send to both, compare for a few weeks, then turn Datadog off.
Move dashboards and monitors. Kloudfuse converts Datadog dashboards and alerts as part of onboarding, and our team works with yours on the migration. Some monitor logic won’t map one-to-one, so plan a manual check. Kloudfuse dashboards are Grafana-compatible. The step-by-step is in our Datadog migration guide.
Know what doesn’t carry over. Integrations that run through Datadog’s own backend instead of the agent, such as webhooks and API-based integrations, won’t follow the agent when you repoint it. List these early.
Datadog alternatives at a glance
| Tool | Best for | Deployment | Pricing unit |
|---|---|---|---|
| Kloudfuse | Data in your cloud, every signal in one place, keeping Datadog Agents | Self-SaaS (managed BYOC) | TB ingested per day |
| groundcover | Kubernetes with eBPF | BYOC or on-prem | Per host |
| Last9 | Single-tenant in your account | Managed in your cloud | Not public |
| Grafana | Teams already on Prometheus and Grafana | SaaS, self-hosted, BYOC (Enterprise) | Platform fee plus per series and per GB |
| SigNoz | Open source on OpenTelemetry | Cloud, self-hosted, BYOC (Enterprise) | Per GB and per million samples |
| New Relic | Small teams with lots of data | SaaS | Per GB plus per user |
| Elastic | Teams already on Elasticsearch | Serverless, hosted, self-managed | Per GB |
| Honeycomb | Debugging high-cardinality data | SaaS, private cloud (Enterprise) | Per event |
| Dynatrace | Enterprises wanting auto-discovery | Check vendor | Per host and per GiB |
| Splunk Observability | Existing Splunk shops | Check vendor | Per host |
Prices from each vendor’s own pricing page, checked October 2026. For a deeper side-by-side, see our comparison pages for Datadog, New Relic, Grafana, Dynatrace, Splunk and Sumo Logic.
Frequently Asked Questions
What is the best alternative to Datadog?
It depends on why you’re leaving. For predictable pricing and data in your own cloud, Kloudfuse. For open source, Grafana or SigNoz. For high-cardinality debugging, Honeycomb.
Why is Datadog so expensive?
Each product has its own meter: hosts, APM hosts, log GB, indexed events and custom metrics. As your environment grows, several meters grow at once, and custom metrics multiply with every new tag combination.
Can I keep the Datadog Agent if I switch?
With some alternatives, yes. Kloudfuse accepts data directly from the Datadog Agent, so you change where it sends data instead of re-instrumenting.
Is there a Datadog alternative that runs in my own cloud?
Yes. Kloudfuse, groundcover and Last9 run in your own cloud account, and SigNoz and Grafana offer BYOC on enterprise plans.
How long does it take to migrate off Datadog?
It depends on how many dashboards and monitors you have. Agents can be repointed quickly; dashboards and monitors take the most time. Running both tools side by side for a few weeks lowers the risk.
Is open-source observability cheaper than Datadog?
The license is cheaper, but you take on the work of running it. Count your team’s time alongside infrastructure costs.
