1. guides
prorm
  • prorm documentation
  • Known test gaps
  • Architecture
  • Core concepts
  • Database types & coverage
  • Getting started
  • Going further
  • compliance
    • Compliance
    • The security toolkit: `prorm audit`
    • Checklists by regulation
    • Declaring personal data
    • Evidence bundles
    • Evidence
    • Frameworks: `@Security`
    • Jurisdictions
    • Platform policies
    • Protecting data
    • The regulation dataset
    • Residency and cross-border transfers
    • Retention & integrity
    • Status
    • Subject rights
  • decorators
    • prorm with decorators
    • Models with decorators
    • Querying a decorated model
    • Relations with decorators
  • diagrams
    • Diagram generators
    • Chen diagram
    • Class diagram
    • Dependency diagram
    • ER diagram
    • Flow diagram
    • From your models to a diagram
    • Gantt diagram
    • Index diagram
    • Migration diagram
    • Model diagram
    • Package diagram
    • Crow’s-foot diagram
    • Schema documentation
    • Sequence diagram
    • State diagram
    • Tree diagram
  • dialects
    • SQL dialects
    • AlloyDB Dialect
    • AlloyDB Omni Dialect
    • Amazon Athena Dialect
    • Aurora DSQL Dialect
    • Amazon Aurora MySQL Dialect
    • Amazon Aurora PostgreSQL Dialect
    • Azure Database for MySQL Dialect
    • Azure Database for PostgreSQL Dialect
    • Azure SQL Database Dialect
    • Azure SQL Managed Instance Dialect
    • Google BigQuery Dialect
    • Citus Dialect
    • ClickHouse dialect
    • Cloud Authentication (`cloud-auth`)
    • Cloud SQL for SQL Server Dialect
    • Cloud SQL for MySQL Dialect
    • Cloud SQL for PostgreSQL Dialect
    • CockroachDB Dialect
    • Azure Cosmos DB for PostgreSQL Dialect
    • CrateDB Dialect
    • Cloudflare D1 Dialect
    • Databricks dialect
    • IBM Db2 Dialect
    • Dolt Dialect
    • Apache Doris Dialect
    • Druid Dialect
    • DuckDB Dialect
    • Exasol dialect
    • Microsoft Fabric Data Warehouse Dialect
    • Firebird Dialect
    • Greenplum Dialect
    • SAP HANA Dialect
    • MySQL HeatWave Dialect
    • Azure Data Explorer (Kusto) Dialect — T-SQL
    • MariaDB Dialect
    • Materialize Dialect
    • MonetDB Dialect
    • MotherDuck Dialect
    • Microsoft SQL Server Dialect
    • MySQL Dialect
    • Neon Dialect
    • OceanBase Dialect
    • openGauss Dialect
    • Oracle Dialect
    • Oracle Autonomous Database Dialect
    • ParadeDB Dialect
    • PolarDB Dialect
    • PostgreSQL Dialect
    • Presto Dialect
    • QuestDB Dialect
    • Aurora RDS Data API Dialect (`rdsdata`)
    • Amazon RDS for Db2 Dialect
    • Amazon RDS for MariaDB Dialect
    • Amazon RDS for SQL Server Dialect
    • Amazon RDS for MySQL Dialect
    • Amazon RDS for Oracle Dialect
    • Amazon RDS for PostgreSQL Dialect
    • Amazon Redshift
    • Amazon Redshift Data API Dialect
    • RisingWave Dialect
    • rqlite Dialect
    • SingleStore Dialect
    • Snowflake Dialect
    • Google Cloud Spanner Dialect
    • Spanner (PostgreSQL interface) Dialect
    • SQLite Dialect
    • StarRocks Dialect
    • Supabase Dialect
    • Azure Synapse Analytics (dedicated SQL pool)
    • Teradata Dialect
    • TiDB Dialect
    • TimescaleDB Dialect
    • Trino Dialect
    • Turso / libSQL Dialect
    • Vertica dialect
    • Vitess Dialect
    • Yellowbrick Dialect
    • YugabyteDB Dialect
  • guides
    • Access rules and behaviour profiles
    • ArangoDB (AQL)
    • Associations
    • Audit levels and escalation
    • Audit logging
    • In the browser
    • Bulk operations
    • Caching
    • Capabilities, tiers and team roles
    • CLI
    • Cloudflare
    • Compliance & data protection
    • Connection Pooling & Performance
    • Security in the console
    • Couchbase: structured N1QL queries
    • Data types
    • Decorators
    • Eager loading
    • Enums and constants
    • Error handling & error types
    • Extensions
    • External fields
    • Foreign data wrappers
    • Graph databases
    • Hooks & lifecycle events
    • Indexes & constraints
    • InfluxDB: structured Flux pipelines
    • klock-errors
    • klock — the `.klock` file
    • Local databases (`prorm db up`)
    • Structured log search — Splunk, Graylog, Papertrail, Sumo Logic
    • Logging
    • Grafana Loki: structured LogQL queries
    • MCP server
    • Migrate — the data transfer subsystem
    • Migrations & Schema Sync
    • Models from an existing database
    • Defining models
    • Multiple databases
    • OrientDB, Neptune and TigerGraph
    • Passwords
    • Importing a Prisma schema
    • Structured PromQL and MetricsQL
    • Query firewall
    • QueryInterface
    • Query operators reference
    • Query optimization
    • Composing find options with `pipe()` and `Q`
    • Query safety: the fail-closed WHERE guard
    • Querying
    • Raw Queries & Query Building
    • Read replicas
    • Schema diffing & migration generation
    • Views, triggers & other schema objects
    • Scopes & default scopes
    • Seeders
    • SPARQL / RDF triplestores
    • SQL constants
    • SQL features
    • SQL function builders
    • SQLite advanced features
    • StoreKey — a prod-orm database as the storage
    • Streaming results
    • Terraform
    • Transactions & Isolation Levels
    • Type mapping: legacy and modern
    • TypeScript types
    • Database users, roles & privileges
    • Validation & constraints
    • Virtual & Computed Fields
    • Writing docs
  • javascript
    • prorm in JavaScript
    • Models in JavaScript
    • Querying in JavaScript
    • Relations in JavaScript
  • runbooks
    • Operational runbooks
    • Connection failures
    • Credential rotation
    • Failover
    • Incident checklist
    • Deploying klock v2
    • Migration rollback
  • running-databases
    • Running databases in Docker
    • Running prorm under Bun in Docker
    • Running Cassandra in Docker with prorm
    • Running ClickHouse in Docker with prorm
    • Running CockroachDB in Docker with prorm
    • Running prorm under Deno in Docker
    • Running Elasticsearch in Docker with prorm
    • Running MariaDB in Docker with prorm
    • Running MongoDB in Docker with prorm
  • prorm control plane
    • Introduction
    • Authentication
    • Errors
    • Getting started
      • Service
        • Is it up
        • This description
    • Account
      • Sign-up and verification
        • Create an account
        • Exchange credentials for a token
        • Verify with the emailed code
        • Verify by clicking the emailed link
        • Name the parent or guardian
        • What a consent link is about
        • Give consent for a child's account
        • Ask the parent or guardian again
        • Send a new code and link
      • Family
        • Everyone this account is responsible for
        • One child, in detail
        • Turn a limit on or off
      • Profile
        • Who am I
        • Change the account
        • Every address on the account
        • Add an address
        • Confirm an address with the code mailed to it
        • Mail a fresh code to an unconfirmed address
        • Make an address primary
        • Remove an address
        • What this caller may do
    • Security
      • Sessions
        • Everywhere this account is signed in
        • Sign out everywhere else
        • Sign one out
        • A token for the CLI on a named machine
        • A token for the customer console, from the staff one
        • Park a sign-in for the CLI to collect
        • The CLI collecting its sign-in
        • The team's members' live sessions
        • Sign a member's session out
        • Sign an account out everywhere
        • Everywhere the account is signed in
        • Sign an account's session out
      • Single sign-on
        • Begin a single sign-on
        • The provider's return
        • Link the signed-in account to a team's provider
        • Which team an address signs in to
        • The team's single sign-on, as a member sees it
        • Set the team's OpenID Connect provider up, or change it
        • Remove the provider
        • The team's domain claims and the DNS record each needs
        • Prove a claimed domain
      • MFA and passwords
        • Start enrolling a second factor
        • Prove the app has it
        • Remove the second factor
        • Ask for a reset link
        • Set a new password with the emailed token
      • Klock
        • Open a device-bound klock session
        • Check in, for the next grant
        • Step up the session asking
        • Rotate the device key
        • Freeze a session
        • Unfreeze a session
        • A guest grant
        • The revocation feed
        • The published keyset
        • What this control plane supports
      • Devices
        • Your devices, or a team's
        • Approve a new device's pairing code
        • One device
        • Rename a device
        • Revoke a device
        • Suspend a device
        • Reinstate a suspended device
        • Wipe a device
        • Ask a device to rotate its key
        • The team's members' devices
        • Revoke a member's device
        • An account's devices
        • Revoke an account's device
        • Wipe an account's device
        • The device fleet
      • Access rules and behaviour
        • The access rules that bear on you
        • Add an access rule
        • This request, as the rules see it
        • One access rule
        • Remove an access rule
        • The behaviour profiles that bear on you
        • Add or change a behaviour profile
        • The profile presets
        • Your behaviour findings
        • Acknowledge a finding
        • One behaviour profile
        • Remove a behaviour profile
        • The team's access rules
        • Add a team access rule
        • Remove a team access rule
        • The team's behaviour profiles
        • Add or change a team behaviour profile
        • Remove a team behaviour profile
        • What the team's profiles caught
        • Acknowledge a team finding
      • Recovery and lockdown
        • Ask for a recovery link
        • Recover access with the emailed code
        • Is the account locked down
        • Lock the account down
        • Release a lockdown
        • Lock a member down
        • Release a member's lockdown
      • MCP tokens
        • Make a delegated token for one MCP client
        • The account's delegated tokens
        • Revoke a delegated token
        • What a delegated token may do now
        • Run one MCP tool with a delegated token
        • Run a device-signed MCP tool call relayed by the MCP Worker
      • Kerberos
        • Give the team its own Kerberos realm
        • The team's realm
        • Remove the team's realm
    • Logging and monitoring
      • Audit log
        • What has been done to the account
        • What has happened to this account
        • The account's audit trail
      • Operations
        • Everything that has been asked of this database
        • One, with its steps
        • Stop it
        • Put a failed one back in the queue
        • Every operation, across all accounts
      • Metrics and events
        • Every measurement, for a graph
        • Who did what to it
    • Billing
      • Plans and pricing
        • What the plans allow
        • The rate card
        • What one database would cost
        • What this account is paying for, itemised
      • Subscriptions and invoices
        • What this month's invoice will say
        • What this account is on
        • What it has been charged
        • Start a Stripe Checkout session
        • Open the customer portal
        • Stripe's events
        • Buy more connection pools
      • Team billing
        • The team's billing account
        • What the team has been charged
        • What this month's invoice will say
        • The team's cards
        • Add a card, through a Stripe-hosted setup session
        • Make a card the team's default
        • Remove a card
        • Start the team paying
        • Open the Stripe portal on the team's customer
        • Change the team's plan
        • Who receives the team's bills
        • How the team pays for transfers
    • Support
      • Support chat
        • My conversations
        • Start a conversation
        • Every support conversation, open first
        • A ticket for the socket
        • One conversation, with who is in it and who is connected
        • Rename a group; open or close a support conversation
        • The lines, oldest first
        • Say something
        • Add somebody
        • Mark read
    • Administration
      • Users
        • Every account
        • One account, with everything hanging off it
        • Grant or revoke admin
        • Verify an account by hand
        • Lock or unlock sign-in
        • Allow or hold provisioning
        • Send the verification mail again
        • Set the plan by hand
      • Platform
        • Every database and store, across all accounts
        • The counts worth putting on a dashboard
      • Staff and permissions
        • What kind of admin the caller is, and which scopes they hold
        • The admin presets
        • Every admin, and the scopes each resolves to
        • Make an account an admin of a kind
        • Change an admin's kind, or a custom admin's scopes
        • Revoke admin
      • Team administration
        • Every team, with its owner and size
        • One team in full — members with their accounts, and projects
        • Hand the team to another member
        • Put an account in the team, active at once
        • Change a member's role
        • Remove a member
      • Assumed identities
        • Assume an identity — act as a customer account
        • The identities the caller holds open
        • End an assumed identity
      • Support inbox
        • The platform's addresses
        • Add a platform address
        • One address, with its counts
        • Change an address's label, routing, or whether it accepts mail
        • Remove an address that holds no mail
        • Make, adopt or repair the address's routing rule
        • Every admin's personal address, made where missing
        • File a raw message, as if it had arrived
        • The inbox
        • One message, whole
        • Mark a message
        • Hand a message to an admin, or take it back
        • Leave an internal note
        • Reply, from the address the message came to
        • What is waiting
    • Databases
      • Databases
        • Data stores
          • Every data store on the account
          • Provision a data store
          • Change a data store that is already running
          • One data store
          • Deprovision it
          • Report what it is using
          • How to reach a data store
        • Connections and credentials
          • The api keys on a database
          • Issue a key
          • Revoke a key
          • The engine users we created
          • Create a user inside the database
          • Give a user a new password
          • Drop a user
          • How the database is reached
          • Set how it is reached
          • Revoke every api key on it
          • How to connect to it
          • The ssh keys that open its bastion
          • Add a key, or have one made
          • The private half of a generated key, once
          • Revoke an ssh key
        • Schema and objects
          • The SQL objects we put in it
          • Create one, and apply it
          • What is in the database, as the engine reports it
          • One table
          • What it would run, without running it
          • Redefine an object in place
          • Drop it
          • The extensions installed in it
        • Migrations
          • A plan compiled to statements, recording nothing
          • The migrations recorded for it
          • Record one
          • Apply everything recorded and not yet run
          • Apply everything pending, oldest version first
        • Lifecycle
          • Grow or shrink it
          • Stop it accepting connections
          • Start it again
          • Move it to another region
        • Backups and replicas
          • Take a backup
          • The backups it has
          • Restore one into it
          • What has been restored into it, and from where
          • Its read copies
          • Add one
        • Policies
          • The catalog
          • One policy
          • What this database is held to
          • Switch a policy on, or change its knobs
          • Switch it off
          • Mark the policies reviewed
          • Evaluate the periodic policies now
          • Who is behind a credential
        • Passwordless
          • The database's passwordless sessions
          • Issue a passwordless session (v1)
          • Revoke a passwordless session
          • Bind a session to a device key
          • Issue a v2 session for a host to bind
          • Redeem an enrolment ticket on the host
          • Check a passwordless credential
          • Run one described query under a passwordless credential
        • Database roles
          • The database's roles
          • Define a role
          • One role
          • Change what a role permits
          • Delete a role, and revoke every key granted it
          • Grant a role, which mints a key
          • Who holds a role
          • The api keys bound to a role
        • Pools
          • A database's connection pools, and how busy each is
          • Open another pool
          • Rename a pool or change its size
          • Close a pool
        • Data plane
          • Run a query against a database
          • Several statements, in order, over one connection
          • Is this database answering
          • What the proxy itself is holding
        • Transfers
          • The strategies this account's tier may use
          • Price a transfer without creating it
          • The account's transfer jobs, newest first
          • Create a transfer job
          • One job, and whether its certificate holds
          • Cancel an unfinished job
          • Start a created job, confirming its price
          • Report a job's status (the runner, or the owner's side of a local run)
        • Every database on the account
        • Provision a database
        • Change a database that is already running
        • One database
        • Deprovision it
        • Read or write it
        • Report what it is using
        • The engines this API will provision
        • Where a managed database can be put
    • Teams
      • Teams
        • Members
          • Who is on it
          • Invite somebody
          • Accept an invitation
          • Change somebody's role
          • Remove somebody
        • Projects
          • The projects a team runs
          • Start one
          • Rename a project, or move its environment
          • Delete a project
          • One project, and what is in it
          • Put a database or store in it
          • What is in a project
          • Take one out
        • The teams this account is on
        • Start one
        • The twelve roles, and what each may do
        • Rename a team
        • Close a team
        • One team
        • Hand the team to somebody else
        • Leave a team
    • Schemas
      • Error
      • ValidationErrors
      • Token
      • PolicyCategory
      • PolicySeverity
      • PolicyAction
      • PolicyTrigger
      • PolicyParam
      • Policy
      • PolicyCatalog
      • PolicyAssignment
      • PolicyOutcome
      • PolicyFiring
      • CredentialTrace
      • Relation
      • Restrictions
      • ChildSummary
      • SignupRequest
      • SigninRequest
      • UpdateMeRequest
      • User
      • Address
      • GeoPoint
      • D1Region
      • Usage
      • UsageReport
      • Resource
      • Database
      • DataStore
      • NewDatabase
      • NewDataStore
      • QueryRequest
      • DatabasePrivilege
      • ApiKey
      • DbUser
      • Access
      • AccessRequest
      • DatabaseObject
      • NewDatabaseObject
      • NewMigration
      • SchemaOperation
      • Migration
      • Operation
      • Session
      • AdminPowers
      • Admin
      • AuditEntry
      • TeamRole
      • Conversation
      • ChatMessage
      • TeamSso
      • Team
      • TeamMember
      • SshKey
      • ConnectionInfo
      • InboxAddress
      • InboxMailbox
      • InboxMessageSummary
      • InboxMessage
      • InboxNote
      • Page
      • Quote
      • Plan
      • ApiError
      • CapabilityRefusal
      • KlockErrorCode
      • PageWindow
      • SessionView
      • KlockStepUpRequest
      • KlockOpenRequest
      • KlockOpenedSession
      • UpdateMeResult
      • CapabilitiesView
      • SessionLimits
      • Device
      • DeviceListed
      • DeviceTeamRef
      • DeviceCommand
      • DeviceAuditLine
      • DeviceActionRequest
      • StaffActionRequest
      • ManagedSession
      • SessionRevoked
      • AccessRule
      • AccessRuleRequest
      • BehaviourProfile
      • BehaviourProfileRequest
      • BehaviourProfileSaved
      • BehaviourFinding
      • BehaviourFindingAck
      • StepUpProof
      • Removed
      • MemberLockdownRequest
      • McpClient
      • McpDelegatedView
      • McpToolCall
      • McpForwardRequest
      • PasswordlessScope
      • RoleScope
      • PasswordlessSession
      • PasswordlessTlsMaterial
      • PasswordlessKerberosMaterial
      • ServiceDeviceKeys
      • ServiceSession
      • DatabaseRole
      • TransferEndpoint
      • TransferJobRequest
      • TransferQuote
      • TransferStrategy
      • TransferJob
      • TransferJobAccess
      • Pool
      • PoolAllowance
      • BillingSubscription
      • Invoice
      • BillingContact
      • PaymentMethod
      • UpcomingCharges
      • TeamBilling
      • TeamKerberosRealm
      • TeamSsoDomain
      • TeamSsoResponse
      • StoreConnection
      • BatchFailure
      • ProxyStatus
  1. guides

Grafana Loki: structured LogQL queries

Source: src/nosql/loki/logql.ts · src/nosql/loki/index.ts · Tests: tests/nosql/loki-logql.test.ts, tests/nosql/loki-endpoints.test.ts

LokiStore originally exposed query(logql) — the query was a string the caller wrote by hand, which is exactly what this library's public API is not supposed to require. This page documents the structured path that closes that gap.

query() has not gone anywhere. It is the escape hatch for what LogQL can express and this compiler deliberately will not, and every refusal below names the feature and points back at it.

Why the spec is not findAll-shaped

Every other store in this library was closed by mapping onto a findAll-like { where, attributes, order, limit } spec, because they answer the question SQL answers: which rows, projected how. LogQL does not. A LogQL query is:

stream selector  →  pipeline  →  (optionally) an aggregation

The pipeline is ordered and stateful. | json invents labels a later | status >= 400 can filter on; | line_format rewrites the line a later |= "x" searches; | drop removes labels a later by (…) would have grouped by. Swap two stages and you get a different answer. A where map is unordered, and limit is a server-side page size rather than a stage — flattening a pipeline into { where, limit, order } would throw away the one property that makes LogQL what it is.

The aggregation half does not return rows at all: it turns logs into a metric time series, which has no findAll counterpart.

So the spec mirrors LogQL's own grammar:

interface LogQuerySpec {
  selector: StreamSelectorSpec;      // { app="api", env=~"prod|staging" }
  pipeline?: LogqlPipelineStage[];   // ordered; order is meaning
}

type LogqlMetricExpr =              // a small expression tree
  | { type: 'logRange'; … }          //   rate({…}[5m])
  | { type: 'unwrappedRange'; … }    //   quantile_over_time(0.99, {…} | unwrap x [5m])
  | { type: 'vectorAgg'; … }         //   sum by (level) (…)
  | { type: 'binary'; … }            //   … / … , … > bool 0.5
  | { type: 'labelReplace'; … }
  | { type: 'vector' | 'literal'; … };

compile(spec) returns { text, kind }, where kind is 'log' or 'metric' — that is what tells the store which result shape to expect.

The surface

MethodEndpointReturns
compileQuery(spec)—{ text, kind }, no I/O
queryLogsRange(spec, opts?)GET /loki/api/v1/query_rangeLokiStreamsResult
queryMetricRange(spec, opts?)GET /loki/api/v1/query_rangeLokiMatrixResult
queryLogsInstant(spec, opts?)GET /loki/api/v1/queryLokiStreamsResult
queryMetricInstant(spec, opts?)GET /loki/api/v1/queryLokiVectorResult | LokiScalarResult
streams(selector?, opts?)GET /loki/api/v1/serieslabel sets
labelValues(name, opts?)GET /loki/api/v1/label/<name>/valueslabel values
query(logql, opts?)GET /loki/api/v1/query_rangeraw escape hatch
import { Op } from 'ts-prorm-orm';
import { LokiStore } from 'ts-prorm-orm/nosql/loki';

const store = new LokiStore({ baseURL: 'http://localhost:3100' });
await store.connect();

const logs = await store.queryLogsRange(
  {
    selector: { where: { app: 'api', env: { [Op.in]: ['prod', 'staging'] } } },
    pipeline: [
      { stage: 'line', op: '|=', value: 'error' },
      { stage: 'json' },
      { stage: 'labelFilter', filter: { kind: 'number', label: 'status', op: '>=', value: 500 } },
    ],
  },
  { since: '1h', limit: 100, direction: 'backward' },
);
{app="api", env=~`(?:prod|staging)`} |= "error" | json | status >= 500

How a query flows

A log query stops after the pipeline and returns streams; a metric query wraps it in a range function and, usually, an aggregation, and returns a matrix or vector:

flowchart LR
  SEL["selector {app=api}: picks streams by indexed labels"] --> LF["line filters |= != |~"]
  LF --> PAR["parser: json / logfmt / regexp"]
  PAR --> LBL["label filters: status >= 500"]
  LBL --> FMT["line_format / label_format / drop"]
  FMT -->|"log query"| STREAMS["streams: [ts, line] per stream"]
  FMT -->|"[5m] range"| RF["rate / count_over_time / unwrap + quantile_over_time"]
  RF --> AGG["sum by / max by / topk"]
  AGG --> MATRIX["matrix (range) or vector (instant)"]

Metric queries, side by side

import { Op } from 'ts-prorm-orm';
import { LokiStore } from 'ts-prorm-orm/nosql/loki';

const store = new LokiStore({ baseURL: 'http://localhost:3100' });
await store.connect();

try {
  // 1. 5xx responses per second, per app
  const errorRate = await store.queryMetricRange(
    {
      type: 'vectorAgg', op: 'sum', by: ['app'],
      expr: {
        type: 'logRange', fn: 'rate', range: '5m',
        log: {
          selector: { where: { env: 'prod' } },
          pipeline: [{ stage: 'json' }, { stage: 'labelFilter', filter: { kind: 'number', label: 'status', op: '>=', value: 500 } }],
        },
      },
    },
    { since: '6h', step: '1m' },
  );

  // 2. p99 request duration per route, from a logfmt field
  const p99 = await store.queryMetricInstant({
    type: 'vectorAgg', op: 'max', by: ['route'],
    expr: {
      type: 'unwrappedRange', fn: 'quantile_over_time', param: 0.99, range: '5m',
      log: { selector: { where: { app: 'api' } }, pipeline: [{ stage: 'logfmt' }] },
      unwrap: { label: 'duration', conversion: 'duration_seconds' },
    },
  });

  // 3. the share of lines that mention an error
  const errorShare = await store.queryMetricInstant({
    type: 'binary', op: '/',
    left: { type: 'vectorAgg', op: 'sum', expr: { type: 'logRange', fn: 'count_over_time', range: '5m', log: { selector: { where: { app: 'api' } }, pipeline: [{ stage: 'line', op: '|=', value: 'error' }] } } },
    right: { type: 'vectorAgg', op: 'sum', expr: { type: 'logRange', fn: 'count_over_time', range: '5m', log: { selector: { where: { app: 'api' } } } } },
  });

  // 4. reshape lines, dropping debug noise
  const lines = await store.queryLogsRange(
    {
      selector: { where: { app: 'api', level: { [Op.ne]: 'debug' } } },
      pipeline: [{ stage: 'json' }, { stage: 'lineFormat', template: '{{.method}} {{.path}} {{.status}}' }],
    },
    { since: '15m', limit: 50 },
  );

  console.log(errorRate.result.length, p99.result, errorShare.result, lines.result.length);
} finally {
  await store.disconnect();
}
# 1.
sum by (app) (rate({env="prod"} | json | status >= 500 [5m]))

# 2.
max by (route) (quantile_over_time(0.99, {app="api"} | logfmt | unwrap duration_seconds(duration) [5m]))

# 3.
sum(count_over_time({app="api"} |= "error" [5m])) / sum(count_over_time({app="api"}[5m]))

# 4.
{app="api", level!="debug"} | json | line_format "{{.method}} {{.path}} {{.status}}"

Two endpoints, two parameter sets

Loki's instant and range endpoints do not take the same parameters, and the ones that do not belong are either rejected or silently ignored — the worse of the two failure modes. The option types are split accordingly, and the store refuses a mismatch instead of sending it:

ParameterInstant (/query)Range (/query_range)
time✅✗ (not in the type)
start / end✗✅
since✗✅ (mutually exclusive with start)
limit✅✅
direction✅✅
step✗✅ metric queries only
interval✗✅ log queries only

step sets the resolution of a returned matrix, so passing it with a log query is refused. interval thins out returned log entries, so passing it with a metric query is refused. Passing start and since is refused too: Loki ignores since when start is set, which hides which window was actually queried.

Two result shapes

The response shape follows the query, not the endpoint:

QueryresultTypeType
log query, either endpointstreams{ stream: Record<string,string>; values: [ts, line][] }[]
metric query, rangematrix{ metric: …; values: [number, string][] }[]
metric query, instantvector{ metric: …; value: [number, string] }[]

Each method is typed for the shape it expects and verifies the resultType it actually received, throwing a DatabaseError on a mismatch rather than handing back a value under a type that is a lie. Passing a metric expression to queryLogsRange() is refused up front, naming queryMetricRange() as the method to use.

Escaping is the whole safety story

A LogQL query travels to Loki as one query= HTTP parameter. There is no wire-level parameter binding — the same situation as the SPARQL dialect, and the reason logql.ts is the only place in this layer where a JavaScript value becomes query text.

The two string forms, and which is emitted where

LogQL has two string literals with different rules, and this compiler uses both on purpose:

PositionForm emittedWhy
Label values (=, !=), line filter text (`=, !=), ip()arguments,json/logfmtexpressions,pattern, line_format/label_formattemplates,label_replace`'s label and replacement argumentsdouble-quoted "…"
Every regular expression: =~, !~ (selector and label filter), |~, !~ (line filter), | regexp, label_replace's pattern, drop/keep matchersbacktick-quoted `…`A raw string: no escape sequences exist inside it at all. \d+ reaches RE2 as \d+ rather than needing \\d+, so there is no double-escaping bug to have.

The cost of the raw form is that a backtick has no representation inside one. A pattern containing a backtick is therefore refused, not re-quoted into the double-quoted form — re-quoting would silently change what every other backslash in the pattern means. Write a literal backtick as \x60 in the pattern; a raw string passes that straight through to RE2.

Validated, not escaped

Identifiers have no quoted form in the LogQL this compiler emits, so they are validated and refused:

  • Label names — stream selector labels, label filter labels, by/without groupings, drop/keep/unwrap targets, json/logfmt expression names, label_format source and destination, template { label } parts, and the label_replace label arguments (quoted, but still validated). Must match [a-zA-Z_][a-zA-Z0-9_]*.
  • Durations ([5m], offset 1h, | latency > 1s, step, since, interval) — whole-number magnitudes with units ns us ms s m h d w y. Fractional durations are refused because Loki's two duration parsers disagree about them.
  • Byte sizes (| size > 20MB) — a number plus one of B kB KiB MB MiB GB GiB TB TiB PB PiB EB EiB.
  • Numbers — must be finite; exponent notation is refused.

Nothing here is rewritten. safeIdent() elsewhere in this repo turns my-label into my_label, which is right for a generated Cypher variable and wrong here: my_label is a different Loki label, so the query would silently read the wrong stream. Loki 3.0 can quote label names inside a stream selector; this compiler does not emit that form, because it would fail against an older server.

Regular expressions

Loki runs Go's regexp (RE2), which has no backtracking, so classic catastrophic-backtracking ReDoS does not apply server-side — and this compiler never executes a caller's pattern itself. What remains is resource abuse and silent mis-parsing, handled structurally:

  • a length cap (4096 characters by default, maxRegexLength to change it);
  • RE2's own {n,m} repeat limit of 1000 — a{1001} is refused here rather than at query time;
  • balanced-paren and terminated-class checks;
  • an outright refusal, naming the construct, of the PCRE features RE2 does not implement: backreferences (\1), lookahead ((?=, (?!), lookbehind ((?<=, (?<!) and atomic groups ((?>).

The pattern is deliberately not compiled with JavaScript's RegExp. The two dialects disagree in both directions: (?P<name>…), which the | regexp parser requires, is a syntax error in JavaScript, while \1 and (?=…) are valid JavaScript and unsupported by RE2. Validating with RegExp would reject exactly the patterns Loki needs.

Coverage

Stream selectors

All four matcher operators, natively:

{ selector: { matchers: [
  { label: 'app',   op: '=',  value: 'api' },
  { label: 'env',   op: '=~', value: 'prod|staging' },
  { label: 'level', op: '!=', value: 'debug' },
  { label: 'host',  op: '!~', value: 'canary-.*' },
] } }
// → {app="api", env=~`prod|staging`, level!="debug", host!~`canary-.*`}

or through the ORM operator surface, folded by the same normalizeWhere() the graph, AQL and N1QL layers use — so there is no second, drifting copy of the operator table:

ORM operatorCompiles toNote
field: 'v', Op.eqfield="v"
Op.nefield!="v"
Op.regexp / Op.notRegexpfield=~… / `field!~`…
Op.in / Op.notInfield=~(?:a|b)``each value regex-escaped; selector regexes are anchored, so this is exact
Op.startsWith / endsWith / substring (+ not*)field=~esc.*`` etc.value regex-escaped
null / Op.isNullfield=""in Loki, ="" matches streams where the label is not set
Op.isNotNullfield!=""

Everything else is refused by name: $or and $not (a selector is a flat conjunction with no alternation or negation), numeric comparisons (selector labels are indexed strings — parse the value out and filter it in the pipeline), and $like/$iLike (translating SQL wildcards would have to guess how an escaped % was meant; use regexp, or the escaped startsWith/endsWith/substring).

Pipeline stages

StageSpecCompiles to
Line filter{ stage: 'line', op: '|=' | '!=' | '|~' | '!~', value | regex | ip }|= "error", |~ \level=\w+`, |= ip("10.0.0.0/8")`
Line filter or chainor: [{ value }, …]|= "a" or "b" — gated, Loki 3.0+
JSON{ stage: 'json', expressions? }| json, | json path="request.path"
logfmt{ stage: 'logfmt', expressions?, strict?, keepEmpty? }| logfmt, | logfmt lvl="level"; flags gated, Loki 3.0+
pattern{ stage: 'pattern', pattern }| pattern "<ip> - <_> <status>"
regexp{ stage: 'regexp', regexp }| regexp \(?P\w+)``
unpack{ stage: 'unpack' }| unpack
decolorize{ stage: 'decolorize' }| decolorize
distinct{ stage: 'distinct', labels }| distinct a, b — gated, Loki 3.x
Label filter{ stage: 'labelFilter', filter }| status >= 400 and level="error"
Label filter from where{ stage: 'where', where }same, via normalizeWhere()
line_format{ stage: 'lineFormat', template | parts }| line_format "{{.status}} {{.path}}"
label_format{ stage: 'labelFormat', formats }| label_format dst=src, other="{{.a}}"
drop / keep{ stage: 'drop' | 'keep', labels }| drop level, method="GET"

pattern is validated (at least one <capture>, no two adjacent captures) and regexp is required to contain a named capture group — both turn a Loki query-time error into a compile-time one.

Templates. line_format and label_format take Go templates, the one place LogQL would otherwise still demand a hand-written string. parts covers the common case structurally:

{ stage: 'lineFormat', parts: [{ label: 'status' }, { text: ' -> ' }, { label: 'path' }] }
// → | line_format "{{.status}} -> {{.path}}"

Literal {{ in a { text } part is emitted as {{"{{"}}, so caller text can never become template code. template stays available for {{ if … }}, the __timestamp__ builtins and sprig functions; it is escaped as a string but its contents are the caller's stated intent, exactly like a regex.

Label filters

LabelFilter is a small expression tree, because LogQL's label filters are one — unlike a stream selector, they nest:

{ kind: 'and', items: [
  { kind: 'number',   label: 'status',  op: '>=', value: 400 },
  { kind: 'or', items: [
    { kind: 'string', label: 'level',   op: '=',  value: 'error' },
    { kind: 'duration', label: 'latency', op: '>', value: '250ms' },
  ] },
] }
// → | (status >= 400 and (level = "error" or latency > 250ms))

Kinds: string, regex, number, duration, bytes, ip, and, or, and where (the ORM surface). Through where, $between becomes a bounded conjunction and $or works — the one place normalizeWhere()'s or survives. $not is refused: LogQL has no negation operator to wrap a group in.

Filtering parser errors is just a label filter on __error__:

{ stage: 'labelFilter', filter: { kind: 'string', label: '__error__', op: '=', value: '' } }

Metric queries

Log range aggregations — rate, count_over_time, bytes_rate, bytes_over_time, absent_over_time:

{ type: 'logRange', fn: 'rate', log: { selector, pipeline }, range: '5m', offset: '1h' }
// → rate({app="api"} |= "error" [5m] offset 1h)

Unwrapped range aggregations — rate, rate_counter, sum_over_time, avg_over_time, max_over_time, min_over_time, first_over_time, last_over_time, stdvar_over_time, stddev_over_time, quantile_over_time:

{
  type: 'unwrappedRange',
  fn: 'quantile_over_time',
  param: 0.99,
  log: { selector, pipeline: [{ stage: 'json' }] },
  unwrap: { label: 'latency', conversion: 'duration' },
  range: '5m',
  by: ['path'],
}
// → quantile_over_time(0.99, {app="api"} | json | unwrap duration(latency) [5m]) by (path)

unwrap.conversion is duration / duration_seconds / bytes; unwrap.filters appends post-unwrap label filters. Grouping is accepted only where Loki accepts it — everything except sum_over_time, absent_over_time, rate and rate_counter; asking for it elsewhere raises LogqlCapabilityError and suggests wrapping the result in a vector aggregation instead.

Vector aggregations — sum, avg, min, max, count, stddev, stdvar, topk, bottomk, plus the gated sort / sort_desc (3.0) and approx_topk (3.3). by / without are mutually exclusive; topk/bottomk require a positive whole param; sort/sort_desc take neither.

{ type: 'vectorAgg', op: 'topk', param: 10,
  expr: { type: 'vectorAgg', op: 'sum', by: ['pod'], expr: rate } }
// → topk(10, sum by (pod) (rate({app="api"}[5m])))

Binary operations — arithmetic + - * / % ^, comparison == != > >= < <= with the optional bool modifier, and logical and / or / unless, between two metric expressions or against a literal. Nested binary operands are parenthesized so precedence follows the spec tree the caller wrote, not the emitted text.

{ type: 'binary', op: '/', left: errorRate, right: totalRate }
// → sum(rate({app="api"} |= "error" [5m])) / sum(rate({app="api"}[5m]))

{ type: 'vector', value: 0 } gives vector(0) for … or vector(0) defaults, and label_replace is available with its regex raw-quoted and both label arguments validated.

Version gating

Where the compiler cannot be sure a construct exists in the server it is talking to, it refuses by name rather than emitting something unverified. Opt in per feature, on the compiler or on the store:

const store = new LokiStore({ baseURL, features: { lineFilterOr: true, logfmtFlags: true } });
FlagConstructIntroduced
lineFilterOr|= "a" or "b"Loki 3.0
logfmtFlags| logfmt --strict --keep-emptyLoki 3.0
sortVectorsort / sort_descLoki 3.0
distinct| distinct a, bLoki 3.x
approxTopkapprox_topk(k, v)Loki 3.3

drop, keep and decolorize (Loki 2.9) are not gated — 2.9 predates every currently-supported release.

What is refused, and why

FeatureReason
$or / $not in a stream selectorA selector is a flat conjunction of matchers; there is no alternation or negation to compile to.
Numeric comparison in a stream selectorSelector labels are indexed strings. Parse the value out (| json, | logfmt, | pattern) and compare it in a label filter.
$like / $iLike anywhereLogQL has no LIKE, and translating one would have to guess how an escaped % or _ was meant. Use regexp, or the escaped startsWith / endsWith / substring.
$not in a label filterLogQL combines label filters with and / or only — there is no group negation.
Ordered comparison against a non-numeric stringLogQL's > >= < <= only apply to numbers, durations and byte sizes.
Vector matching modifiers (on, ignoring, group_left, group_right)Not offered: the compiler cannot verify their behaviour across Loki versions, and emitting an unverified modifier silently changes which series join.
The @ modifierNot part of LogQL.
Quoted label names in a selector (Loki 3.0)Would parse-error against an older server; an invalid label name is refused instead.
PromQL's quantile vector aggregationDoes not exist in LogQL — use quantile_over_time on an unwrapped range.
A backtick in a regular expressionThe raw-string form has no escape for it. Use \x60 inside the pattern.
An empty stream selectorLoki refuses it; it would mean "every stream in the cluster".
An empty where on a label filter stageIt would emit a filter that matches everything. Drop the stage instead.

Each of these throws LogqlCapabilityError (a DatabaseError subclass carrying feature) or ValidationError, both naming what was refused and pointing at LokiStore.query() where a hand-written query is genuinely the answer.

Verification status

Unit / mock-verified only. The compiler tests are pure string comparisons — every construct asserts its exact compiled LogQL, and a hostile-input suite pushes a quote-breakout payload through every value-carrying position (label value, line filter, label filter, parser expression, pattern, template, drop matcher, label_replace replacement) and pins the whole output, plus a backtick-breakout attempt through every regex position and a quote-breakout attempt through every identifier position. The endpoint tests inject a mock LokiHttpClient and assert the path and query parameters.

There is no live Loki server in the test run. Query syntax is verified against Loki's documented LogQL grammar, not by execution.

Related reading

  • LokiStore — the store, its connection options and its ingest/metadata methods
  • SPARQL / RDF triplestores — the other language here with no wire-level parameter binding
  • Couchbase (N1QL) — the compile-only-module precedent this follows
  • Query operators — the Op.* surface normalizeWhere() folds
Modified at 2026-10-04 00:40:16
Previous
Logging
Next
MCP server
Built with