Why engineering teams choose Last9 vs groundcover
Both run in a cloud account you own, and both have an AI agent and an MCP server. Through Last9's MCP server an agent adds drop rules and edits dashboards, and the Last9 agent runs on your model provider. groundcover's agent runs on Claude, and its eBPF sensor collects with no code change. Last9 queries in PromQL and LogQL, where groundcover adds gcQL, its own language.
100+ native integrations, OpenTelemetry-first. A Last9 engineer runs the move with your team.
Named a Gartner® Cool Vendor in AI for SRE and Observability, cited for its unified telemetry platform.
The Last9 agent runs on your model provider, with no per-investigation charge, in your own cloud account.
Through the hosted Last9 MCP server, an agent adds drop rules and edits dashboards.
At a glance
Last9 vs groundcover at a glance
The questions engineering teams ask when they compare the two. groundcover's BYOC deployment is the product compared here, with its on-prem option named where deployment differs.
Does it have an MCP server for coding agents?
Yes, hosted, since March 2025, for any MCP client. An agent can act through it: add drop rules and edit dashboards
Yes, since May 2025, marked work in progress. Its tools can only query data and draft dashboards or monitors, which you upload yourself
Can the AI use our model provider, in our own cloud?
Yes, any model provider or inference endpoint. The Last9 agent runs inside your own cloud account and picks the model for each step
On Claude only. Agent Mode runs Anthropic's Claude through Bedrock, Vertex AI, or Microsoft Foundry in your cloud, or on your own Anthropic key
Can it show what our coding agents cost?
Yes, in one view for Claude Code, Cursor, Copilot, and Codex, with cost per developer and week-over-week change
One dashboard per tool. Claude Code sends logs and metrics, and Codex sends logs only
Which query languages does it use?
PromQL and LogQL, or none: build log queries in the query builder, or ask in plain English
PromQL or MetricsQL for metrics. Logs, traces, and Kubernetes events use gcQL, groundcover's own language, and so do the dashboards and monitors built on them
How much metric cardinality can I keep?
No hard limit. 20M series per metric per day by default, raised on request, with no sampling and no charge per label or metric name
No per-series charge. Metrics its sensor scrapes default to 1M active series an hour and 5M a day, raised in configuration
Can I filter telemetry before I pay to store it?
Yes, for metrics, logs, and traces, with no code to write. Drop, remap, redact, forward, and aggregate them at ingest, from the UI or Terraform
Yes, in a pipeline per signal that admins edit: drop, parse, and mask logs and traces, turn them into metrics, and aggregate metrics
How is it priced?
One licence. The infrastructure runs on your own cloud bill, at your negotiated rates
Per node or host, with every signal in the host fee and no metering of data volume. On BYOC, the infrastructure is on your own cloud bill
Can it run in our own cloud account?
Yes. Last9 runs in your own cloud account, on AWS or GCP in any region, and manages it for you, including the AI
Yes, on AWS, GCP, or Azure, managed by groundcover, with the web app and sign-in in groundcover's cloud. On-prem and air-gapped options need Enterprise
How do services get instrumented?
No code change for Java, Node.js, Python, and .NET on Kubernetes: one annotation per deployment. Other workloads send OpenTelemetry or Prometheus, with 100+ native integrations
An eBPF sensor collects traces, logs, and metrics with no code change, in any language, on Kubernetes and Linux hosts. It also takes OpenTelemetry
| Capability | Last9 | groundcover |
|---|---|---|
| Does it have an MCP server for coding agents? | Yes, hosted, since March 2025, for any MCP client. An agent can act through it: add drop rules and edit dashboards | Yes, since May 2025, marked work in progress. Its tools can only query data and draft dashboards or monitors, which you upload yourself |
| Can the AI use our model provider, in our own cloud? | Yes, any model provider or inference endpoint. The Last9 agent runs inside your own cloud account and picks the model for each step | On Claude only. Agent Mode runs Anthropic's Claude through Bedrock, Vertex AI, or Microsoft Foundry in your cloud, or on your own Anthropic key |
| Can it show what our coding agents cost? | Yes, in one view for Claude Code, Cursor, Copilot, and Codex, with cost per developer and week-over-week change | One dashboard per tool. Claude Code sends logs and metrics, and Codex sends logs only |
| Which query languages does it use? | PromQL and LogQL, or none: build log queries in the query builder, or ask in plain English | PromQL or MetricsQL for metrics. Logs, traces, and Kubernetes events use gcQL, groundcover's own language, and so do the dashboards and monitors built on them |
| How much metric cardinality can I keep? | No hard limit. 20M series per metric per day by default, raised on request, with no sampling and no charge per label or metric name | No per-series charge. Metrics its sensor scrapes default to 1M active series an hour and 5M a day, raised in configuration |
| Can I filter telemetry before I pay to store it? | Yes, for metrics, logs, and traces, with no code to write. Drop, remap, redact, forward, and aggregate them at ingest, from the UI or Terraform | Yes, in a pipeline per signal that admins edit: drop, parse, and mask logs and traces, turn them into metrics, and aggregate metrics |
| How is it priced? | One licence. The infrastructure runs on your own cloud bill, at your negotiated rates | Per node or host, with every signal in the host fee and no metering of data volume. On BYOC, the infrastructure is on your own cloud bill |
| Can it run in our own cloud account? | Yes. Last9 runs in your own cloud account, on AWS or GCP in any region, and manages it for you, including the AI | Yes, on AWS, GCP, or Azure, managed by groundcover, with the web app and sign-in in groundcover's cloud. On-prem and air-gapped options need Enterprise |
| How do services get instrumented? | No code change for Java, Node.js, Python, and .NET on Kubernetes: one annotation per deployment. Other workloads send OpenTelemetry or Prometheus, with 100+ native integrations | An eBPF sensor collects traces, logs, and metrics with no code change, in any language, on Kubernetes and Linux hosts. It also takes OpenTelemetry |
Differences
Where Last9 and groundcover differ
Three places the two platforms take a different approach, with the tradeoffs stated.
AI on your model provider, with a write path
-
Through the hosted MCP server, an agent adds a drop rule and stops ingesting the noisy source it found, or edits the dashboard it used. It works with any MCP client.
-
The production agent, Last9, responds to pages with context from Slack, Linear, and PagerDuty. It runs on your model provider or inference endpoint and picks the model for each step. There is no per-query or per-investigation charge, and customer data is never used for training.
-
Click a point on a chart, a table row, an alert, or an exception, and choose Ask Last9 AI. The investigation starts from that value, in its service context.
-
The Coding Agents view puts Claude Code, Cursor, Copilot, and Codex side by side, with cost per developer, outliers, and week-over-week change. Agents Monitoring covers your own LLM apps on OpenTelemetry GenAI conventions.
groundcover
Agent Mode runs on Anthropic's Claude in your cloud, with spending caps, and the Slack agent is on BYOC only. Through connectors it opens pull requests and updates PagerDuty and Jira. Its MCP server queries data and drafts dashboards and monitors, and its docs state that the agent tools do not edit monitors, change pipelines, or persist dashboards.
Open query languages, and control at ingest
-
PromQL and LogQL natively, with a builder and plain-English Ask Mode for anyone who does not write queries. Your own Grafana can query Last9 metrics as a Prometheus data source.
-
Drop, remap, redact, forward, and aggregate metrics, logs, and traces at ingest, from the UI or as code with the Terraform provider. Every rule is previewed against live data before you save it.
-
Labels like customer_id or tenant stay on the metric, with no charge per label or metric name.
groundcover
Metrics answer PromQL or MetricsQL. Logs, traces, and Kubernetes events use gcQL, a pipe language groundcover built, and dashboards and monitors on those signals are written in it. SQL monitors reach the ClickHouse store underneath. Admins edit a pipeline per signal, and the sensor samples traces.
Your cloud account, one licence, and engineers who know it
-
Last9 runs single-tenant in your own cloud account, on AWS or GCP in any region, and manages it for you, including the AI. Last9 sizes the cluster, tunes it, and runs the upgrades.
-
One licence covers distributed tracing, profiling, custom metrics, custom dashboards, alerting, RBAC, and forward deployed engineers. The infrastructure runs on your own cloud bill, at your negotiated rates.
-
Support is a shared channel in Slack, Microsoft Teams, or Google Chat, with forward deployed engineers who know your setup.
groundcover
BYOC runs the backend in your account on AWS, GCP, or Azure, managed by groundcover, while the web app and sign-in stay in groundcover's cloud. Pricing is per node or host, with RBAC on Enterprise. On-prem and air-gapped deployments, which your team runs, need Enterprise.
Which fits you
Which one fits your team?
Choose Last9 if
- You want an MCP server that lets an agent add drop rules and edit dashboards, not only query and draft
- You want the AI agent on your own model provider, not on one vendor's model
- You want PromQL and LogQL, or plain English, instead of gcQL
- You want one view of what Claude Code, Cursor, Copilot, and Codex cost, per developer
- You want OpenTelemetry instrumentation that moves with you, and engineers in a shared channel who run the move
Choose groundcover if
- Most of your services are uninstrumented, including Go, and you want telemetry with no code change
- You need the backend on Azure, or fully on your own premises
- You want LLM calls captured without an SDK
- You want a free plan to try it on before you buy
Migration
Moving off groundcover
What the switch involves, and what Last9 does for you.
Forward deployed engineers
A Last9 engineer runs the move with your team.
Talk to an engineerInstrument once, with OpenTelemetry
Services already sending OTLP to groundcover need only a second exporter. Services the sensor collects need OpenTelemetry: on Kubernetes, one annotation instruments Java, Node.js, Python, and .NET with no code change.
See the Kubernetes operatorQueries without gcQL
A Last9 engineer moves your dashboards and alerts with your team, and rewrites gcQL queries in PromQL or LogQL. PromQL metric queries carry over, and anyone who does not write queries uses the builder or plain-English Ask Mode.
Run both
Keep the groundcover sensor running and point an OpenTelemetry Collector at both platforms. Compare them on your own incidents before you cut anything over.
Objections
The objections we hear most
Each one is a fair reason to hesitate.
groundcover already runs in our cloud and needs no code changes.
Last9 runs in your own cloud account too, and manages it for you, including the AI. On Kubernetes, one annotation per deployment instruments Java, Node.js, Python, and .NET with no code change. What you gain is OpenTelemetry instrumentation that any backend reads, and PromQL and LogQL instead of gcQL.
groundcover already has Agent Mode and an MCP server.
Agent Mode runs on Claude only, and the groundcover MCP server queries data and drafts dashboards and monitors for you to upload. The Last9 agent runs on your model provider or inference endpoint, in your own cloud account, with no per-investigation charge, and through the hosted Last9 MCP server an agent adds drop rules and edits dashboards.
groundcover's support answers us in Slack.
So does Last9's. Support is a shared channel in Slack, Microsoft Teams, or Google Chat, and the forward deployed engineers in it know your setup and run the move with your team.
Our services have no instrumentation.
On Kubernetes, the Last9 operator instruments Java, Node.js, Python, and .NET with one annotation per deployment. Go and other languages need an OpenTelemetry SDK. Until they have one, keep the groundcover sensor on those services and send the rest to Last9.
Is Last9 proven at our scale?
Last9 regularly handles 500M+ events a minute and 60M+ concurrent users, and has monitored 12 of the world's 20 largest streaming events. It is SOC 2 Type II certified and PCI ready, with SSO and access controls. Teams across fintech, healthcare, gaming, media, and commerce run it in production, including Replit, Pine Labs, Tata 1mg, and Housing.com.
67%
Average reduction in observability costs
2.5x
More telemetry retained without sampling
100+
Native integrations, OpenTelemetry-first
24x7
Support from forward deployed engineers
Customers
What do Last9 customers say?
With Last9, we just eliminated the toil. No more worrying about failing dashboards or alerts — we can finally use our metrics. It just works.
Last9 allowed us to offload the operational overhead of scaling so we could focus on business alerting rather than infrastructure management.
Frequently asked questions
Ask AI about this comparison
Do not take our word for it. Send the question to an assistant and check what it finds.
Can't find the answer to your question?
Talk to our support team for help.
Start observing for free. No lock-in.
OpenTelemetry · Prometheus
Just update your config. Start seeing data on Last9 in seconds.
Datadog · New Relic · Others
We've got you covered. Bring over your dashboards & alerts in one click.
Built on Open Standards
100+ integrations. OTel native, works with your existing stack.
