Pricing in detail
KubeBolt Cloud plans, usage units and the AI credit model. We bill for infrastructure, not people.
During launch we bill only the plan. If you exceed your limits, the overage is measured but not billed yet. Team is live today; when we enable Business we'll decide whether to start charging overage — always with prior notice and a grace period. User limits marked soft don't block: if you exceed them, we suggest a higher plan.
| Plan comparison | Free Free | Team Recommended | Business Coming soon | Enterprise Contract |
|---|---|---|---|---|
| Included limits | ||||
| Price / mo | $0 €0 | $49 €45 | $99 €92 | Custom |
| Clusters | 2 | unlimited | unlimited | unlimited |
| Nodes / pods | 10 / 150 | usage-based | usage-based | negotiated |
| Pods included per node | hard cap | 50 | 80 | 150 |
| Users | 3 (hard cap) | 15 (soft) | unlimited | unlimited |
| Retention | 15 days | 30 days | 90 days | 365 days |
| Active series included | 25,000 | 150,000 | 750,000 | 3,000,000 |
| AI credits / mo | 100 | 1,500 | 3,600 | 50,000 |
| Usage units Measured, not billed during launch. These are the prices that will apply once overage billing is switched on. | ||||
| Node-hour production | hard cap | $0.012 €0.011 | $0.018 €0.017 | negotiated |
| Node-hour non-production | hard cap | $0.003 €0.0028 | $0.004 €0.0037 | negotiated |
| Pod-hour above the allowance | hard cap | $0.0008 €0.00074 | $0.0006 €0.00056 | negotiated |
| Pod overage ceiling | - | 50% of the node charge | 50% of the node charge | 50% of the node charge |
| Extra series / 1K | hard cap | $4.00 €3.70 | $3.50 €3.25 | negotiated |
| Extra AI credit | hard cap | $0.03 €0.028 | $0.025 €0.023 | negotiated |
| Capabilities | ||||
| OAuth / SSO | - | ✓ | ✓ | SSO/SAML |
| Autopilot | Beta | ✓ | ✓ | full |
| Support | Community | Email 8×5 | Chat & email | 24×7 · dedicated CSM |
During launch the plan fee is billed and nothing else.
The day it switches on there will be notice and a grace period. Until then this is metered and never reaches your invoice.
| Line | Billable volume | Amount |
|---|
An estimate, not a quote. Based on 720 hours a month. AI credits are the only approximate line: 50 to 100 per incident and 4 to 31 per conversation, measured in production. Everything else is deterministic.
AI credit model
The credit is elastic: it measures real AI work. A full incident, from triage to postmortem, costs very little. The Insights Engine detectors are deterministic and consume no credits.
The credit measures real compute. A short query costs little; an Autopilot incident, roughly 50 to 100 credits.
Every plan includes credits, with a live meter of what's used. KubeBolt warns you in the product before the allowance runs out, and emails your organization's admins once a month — so whoever can actually do something about it finds out in time. Deterministic rules consume no credits.
The LLM diagnoses freely, but only applies actions from an audited catalog. It's not token savings: it's what makes it deployable in production.
Comparison
| Product | Cost per incident | Executes? | How it's billed |
|---|---|---|---|
| KubeBolt | $1.50–$3 | ✓ Yes | 50–100 credits, closed range |
| Dash0 Agent0 | $3–$6 | No | 5–10 credits × $0.60, drafts a PR |
| AWS DevOps Agent | ~$2.50–$5 | No | $0.0083/agent-second, you pay the clock |
| Datadog Bits AI | $6.50–$8.45 | No | ~6.5 AI credits × $1–$1.30 |
Dash0 Agent0 investigates and drafts a PR, but a human merges it; by default it never touches production. KubeBolt applies the remediation from an audited catalog.
AWS bills per agent-second ($0.50/min): a long incident scales without warning until it's over. And it only recommends — a human approves and applies the fix.
Datadog Bits AI investigates and suggests, it doesn't execute. Its per-investigation cost (AI credits) is several times KubeBolt's — which also remediates.
The infrastructure differentiator
Grafana Cloud bills four meters separately: host hours, container hours, active series and the per-user AI Assistant. Here's the same cluster on both, meter by meter: 20 production nodes, 5 non-production, 175 pods, 50,000 series and 12 SREs.
They meter compute twice, per node and per container. We meter it once: the node already carries 50 pods, so the fair comparison is the sum. On the bare node row they are cheaper, but summed their compute is $219 and ours is $184. From there the gap only widens, because your 50,000 series are already in the plan and their Assistant is billed per head.
Two of their four meters are zero on KubeBolt.
| Meter | Grafana | KubeBolt |
|---|---|---|
| Platform / plan | $19 | $49 |
| Nodes · 18,000 host hours | $158 | $184 |
| Pods · 126,000 container hours | $62 | inside the node |
| Compute subtotal | $219 | $184 |
| Active series · 50K | $195 | included |
| AI · Assistant / credits | $180 | $124 |
| Total per month | $613 | $356 · -42% |
The scale-up profile, the same one the calculator opens on: 20 production nodes, 5 non-production, 175 pods, 50,000 active series and 12 SREs. Compared against the Team plan because that is what you can buy today; the gap widens with scale, and the same exercise on a 120-node cluster lands at -52% on Business. The Grafana column comes from their own calculator (Cloud Pro, Aug 2026 list) with the 20% adaptive reduction applied to series and one container per pod. The two compute rows are the subtotal's breakdown, not separate line items. Both sides use the same 720 h/month base. The KubeBolt column comes from the same rate card as the calculator above, Team plan.
Frequently asked questions
No. KubeBolt bills for monitored infrastructure: nodes, series and AI credits. Free is capped at 3 users (hard cap) and Team at 15 (soft cap: if you exceed it we suggest a higher plan, it doesn't block); Business and Enterprise have no per-person cap.
During launch, nothing: overage is measured but not billed. Team is live today. When we enable Business we'll decide whether to start charging overage, with prior notice and a grace period.
The credit measures real AI work, proportional to compute. As a rough reference, a full Autopilot incident tends to run around 50 to 100 credits, and a Copilot query, 4 to 16. These are estimates, not fixed figures.
Both — use the selector above. Each plan and each usage unit has its own published amount in each currency, with no automatic conversion.
A series with at least one sample in the last 15 minutes — a shorter window than Grafana Cloud's 20 minutes, so fewer of your series count as active. We read your cardinality every 5 minutes, keep each hour's high point, and bill the 95th percentile of those hourly readings across the month: a spike during a rollout doesn't set the price of your whole month.
The OSS (Apache 2.0) is available to everyone. The mid tiers live in Cloud because operating the infrastructure is part of the product. Self-Hosted Enterprise is for compliance that requires on-premise.
Every node carries included pods — 50 on Team, 80 on Business, 150 on Enterprise — pooled across the account, so a dense node is covered by an empty one. You only pay for pods above that pool, and that charge never exceeds 50% of what you pay for nodes, however dense your cluster gets. At the usual 10 to 25 pods per node you never reach it.
You're billed for each active node's real hours: a node that exists for 10 days of the month pays 10 days, not the whole month. Autoscaling and powering down environments out of hours cuts the bill directly, and a non-production node-hour costs considerably less than a production one.
No — included credits reset monthly and don't roll over. That's why the plans carry generous bundles rather than selling you packs that expire. You can see live consumption inside the product, and we warn you before they run out.
Nothing is billed — it stops. At the series limit the agent stops sending, and once credits run out the AI stops answering until next month. On Free there's no overage to charge, so there's no surprise invoice either.
Built for LATAM
off the plan. Full predictability for your finances.
the first year, for companies under 2M ARR and under 50 employees.
Billing in MXN, COP, CLP, ARS and PEN via Stripe.
Guarantee
Cancel your subscription within the first 30 days of purchase and get a 100% refund. No hidden conditions.
No card. 2 clusters, 10 nodes and 100 AI credits per month to start.
Start free →The KubeBolt AI SRE · visiting
Kobi can be wrong. Official pricing lives in the plans section.