Observability

Ask in English, get an answer in SQL,
one query language for every signal

NovaQL is the unified query layer over logs, metrics, traces, and incidents. Type a question in plain English and NovaQL transpiles it into a precise query you can read, edit, save, and share. No more "wait, was that LogQL or PromQL again?"

Get Started Talk to Sales
app.novaaiops.com / novaql
● LIVE
4
Signal types in one query
< 200ms
Median query latency
AI
Plain-English transpile
Save
Per-team query library
One Language, Four Signals

Logs, metrics, traces, incidents, one query

NovaQL lets you join across signal types in a single query. Want logs from the same trace as a slow request? Want metrics from the same service as an active incident? It is one SELECT with a JOIN. The engine pushes down to the right backend (Loki, Prometheus, Tempo, Postgres) automatically.

  • ✓
    Cross-signal joins: JOIN logs ON traces.trace_id = logs.trace_id, works as a first-class operation
  • ✓
    Pushdown planner: planner picks the right backend per subquery so you do not eat 10x the storage cost
  • ✓
    Schema introspection: tab-complete column names per signal type; the schema updates as new sources connect
app.novaaiops.com / novaql · join
Plain-English Mode

Type the question, read the query

Plain-English mode uses the AI fleet to transpile your sentence into NovaQL syntax. The transpilation always shows the generated query so you can verify it before running. If it is wrong, you fix the SQL, you are never stuck with a "computer says no" output.

  • ✓
    Always shows the SQL: no hidden query, the transpilation is visible, editable, and saved alongside the English text
  • ✓
    Schema-aware: transpiler knows your tenant's schema so it does not invent column names that do not exist
  • ✓
    Confidence flag: each transpilation reports a confidence score; low scores prompt for clarification before running
app.novaaiops.com / novaql · english
Saved Queries & Library

Build a team-wide query library

Save any query with a name, tags, and a description. Saved queries show up in autocomplete and on the team query library page. New on-call engineers can run "the right query" without inventing it from scratch. Queries are versioned, so a refactor never breaks an older saved view.

  • ✓
    Per-team library: queries scoped to the team; each library entry has tags, description, and last-edit timestamp
  • ✓
    Versioned forever: every save creates a new version; older views keep working even if someone refactors the query
  • ✓
    Permissioned: queries respect tenant scoping and role; junior on-call cannot run a query that joins to billing data
app.novaaiops.com / novaql · library
Cost & Quota

Queries cost real money, see the cost first

Every NovaQL query is cost-estimated before it runs. Big scans get a confirmation prompt. Each tenant gets a query-cost budget per day, configurable in Quota Cost Control. Long-running queries are cancellable. Saved queries that run via cron count against the budget too, with attribution to the schedule.

  • ✓
    Pre-flight cost estimate: rows scanned, GB read, est latency, shown before you press run
  • ✓
    Soft and hard caps: soft cap warns the user; hard cap rejects the query and writes a runbook to up the budget
  • ✓
    Attribution: cost reports show which queries, which users, and which scheduled jobs ate the budget
app.novaaiops.com / novaql · cost
Video walkthrough coming soon

Subscribe to Nova AI Ops on YouTube for demos, tutorials, and feature deep-dives.

Stop context-switching between query languages

PromQL, LogQL, TraceQL, SQL, NovaQL covers them all. Plain English when you are tired, real syntax when you are not.

Get Started Request a Demo