sketchlog

SketchLog

Bounded-memory telemetry sketches for streaming metrics, live analytics, and production operations.

Status Python Version License

SketchLog compresses high-throughput telemetry streams into bounded-memory, mergeable statistical summaries. It combines DDSketch for latency percentiles, HyperLogLog for cardinality, and Count-Min Sketch for event frequencies, then adds production surfaces for APIs, dashboards, alerts, exporters, Kubernetes, and multi-language SDKs.

The goal is simple: keep the operational signal without retaining every raw event.

Why SketchLog

Bounded memory per stream instead of unbounded raw-event storage.
Explicit approximation guarantees for percentiles, cardinality, and frequency.
Mergeable sketch state for distributed systems and edge deployments.
Live query, dashboard, SLO, anomaly, and export workflows built on the same compact telemetry model.
Production-minded release engineering with CI, coverage, security scanning, docs, Helm packaging, container checks, and public demo verification.

SketchLog is currently a production-minded open-source beta. The core data structures and release pipeline are heavily tested, while some higher-level operational features continue to mature.

Live demo

Open the hosted playground: https://sbalavignesh123.github.io/sketchlog/demo/

Run the deterministic product demo and its end-to-end verifier:

docker compose -f demo/compose.yml up --build --wait

Open http://localhost:4173.

The demo includes live ingestion, percentile sketches, cardinality metrics, Streaming SQL examples, anomaly evidence, tenant isolation, exporter examples, and operational readiness checks. See the demo runbook for ports, cleanup, verification, and recording guidance.

Core capabilities

Area	Capabilities
Sketch engine	DDSketch, HyperLogLog, Count-Min Sketch, bounded sparse stores, merge contracts
Server	FastAPI service, HTTP ingestion, WebSocket updates, health/readiness, OpenAPI contract
Streaming analytics	Streaming SQL, query builder, sketch diffing, anomaly comparison, Smart SLO engine
Multi-tenancy	Namespace quotas, namespace-scoped auth, RBAC checks, tenant-safe metrics
Distributed mode	Authenticated Sketch Mesh, versioned tombstones, peer allowlists, convergence tests
Dashboards	React dashboard SDK, standalone demo dashboard, Grafana dashboard, Grafana datasource plugin
Integrations	Prometheus, OpenTelemetry, OpenTelemetry Collector component, Loki, Datadog, New Relic
Clients	Python embedded API, Python async client, TypeScript client, Go HTTP client, native Go sketches
Deployment	Docker image, Docker Compose demo, Helm chart, Kubernetes Operator manifests
Operations	Doctor checks, alert manager, rate limiting, TLS/mTLS helpers, DB hardening, benchmark lab
Runtime targets	Python, C++, TypeScript, Go, WebAssembly, Linux eBPF collector

Architecture

flowchart LR
    subgraph Producers
        Python[Python SDK]
        TypeScript[TypeScript SDK]
        Go[Go SDK]
        OTel[OpenTelemetry]
        EBPF[Linux eBPF]
    end

    Python --> API[SketchLog API]
    TypeScript --> API
    Go --> API
    OTel --> API
    EBPF --> API

    API --> Registry[Namespace registry]
    Registry --> Stream[Bounded StreamLog]
    Stream --> DDS[DDSketch]
    Stream --> HLL[HyperLogLog]
    Stream --> CMS[Count-Min Sketch]
    Stream --> Analytics[SQL, SLO, diff, anomaly]
    Stream <--> Mesh[Authenticated Sketch Mesh]
    API --> Dashboard[Live dashboards]
    API --> Exporters[Prometheus, OTel, Loki, Datadog, New Relic]

Each stream stores compact sketch state rather than raw telemetry. See the architecture guide for merge behavior, windowing, drift detection, memory limits, and runtime details.

Installation

Install the embedded Python library:

pip install sketchlog

Install the standalone server:

pip install "sketchlog[server]"
sketchlog-server --host 127.0.0.1 --port 8000

Install the TypeScript client:

npm install @sketchlog/client

Install the Go client:

go get github.com/SBALAVIGNESH123/sketchlog/clients/go@v1.2.4

Quickstart

Python embedded sketch

from sketchlog import StreamLog

log = StreamLog()
log.add_latency(42.5)
log.add_batch([15.0, 88.2, 42.1])
log.add_unique("user_12345")
log.add_event("cache_miss", count=5)

print(f"p99 latency: {log.p99():.2f} ms")

Python async client

from sketchlog.async_client import AsyncSketchLogClient

async with AsyncSketchLogClient("http://localhost:8000") as client:
    await client.ingest_events(
        namespace="production",
        stream="api.latency",
        latencies=[42.5, 15.0, 88.2],
        uniques=["user_12345"],
        events={"cache_miss": 5},
    )

TypeScript client

import { SketchLogClient } from '@sketchlog/client';

const client = new SketchLogClient({ endpoint: 'http://localhost:8000' });

await client.ingestEvents('production_api', {
  latencies: [42.5, 15.0, 88.2, 42.1],
  uniques: ['user_12345'],
  events: { cache_miss: 5 },
});

Go client

import "github.com/SBALAVIGNESH123/sketchlog/clients/go"

client := sketchlog.NewClient(sketchlog.ClientOptions{
    Endpoint: "http://localhost:8000",
})

batch := sketchlog.EventBatch{
    Latencies: []float64{42.5, 15.0, 88.2, 42.1},
    Uniques:   []string{"user_12345"},
    Events:    map[string]int64{"cache_miss": 5},
}

err := client.IngestEvents(ctx, "production_api", batch)

Deployment

Run the server with Docker:

docker run --rm -p 8000:8000 ghcr.io/sbalavignesh123/sketchlog:1.2.4

Install with Helm:

helm upgrade --install sketchlog oci://ghcr.io/sbalavignesh123/charts/sketchlog \
  --version 1.2.4

Run mesh mode on Kubernetes:

helm upgrade --install sketchlog ./charts/sketchlog \
  --set replicaCount=3 \
  --set mesh.enabled=true \
  --set-string mesh.clusterSecret="$CLUSTER_SECRET"

SketchLog also includes Kubernetes Operator manifests and documentation for declarative cluster management.

Documentation

The full documentation is published at sbalavignesh123.github.io/sketchlog.

Important sections:

Build docs locally:

pip install ".[docs]"
mkdocs serve

Development

Install the development environment:

make dev-install

Run checks:

make test
make test-go
make test-ts
make docs

Run the full demo:

make demo

What SketchLog is not

SketchLog is a streaming metrics compression layer. It is deliberately not:

A tracing system. It does not store request paths, spans, correlation IDs, or causal chains.
A time-series database. It does not provide raw historical drill-down or full label indexing.
A raw log storage system. It keeps compact summaries, not every event payload.
An exact analytics engine. Results are probabilistic with documented error bounds.

Community

GitHub: https://github.com/SBALAVIGNESH123/sketchlog
Documentation: https://sbalavignesh123.github.io/sketchlog/
Slack: https://join.slack.com/t/sketchlog/shared_invite/zt-41kc03dnl-tiyHm4Gr2CbaJWuGHxdbiQ

License

SketchLog is released under the MIT License.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

sketchlog-1.2.4.tar.gz (267.7 kB view details)

Uploaded Jul 8, 2026 Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

sketchlog-1.2.4-cp314-cp314-win_amd64.whl (504.7 kB view details)

Uploaded Jul 8, 2026 CPython 3.14Windows x86-64

sketchlog-1.2.4-cp314-cp314-musllinux_1_2_x86_64.whl (1.4 MB view details)

Uploaded Jul 8, 2026 CPython 3.14musllinux: musl 1.2+ x86-64

sketchlog-1.2.4-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (341.0 kB view details)

Uploaded Jul 8, 2026 CPython 3.14manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

sketchlog-1.2.4-cp314-cp314-macosx_11_0_arm64.whl (315.0 kB view details)

Uploaded Jul 8, 2026 CPython 3.14macOS 11.0+ ARM64

sketchlog-1.2.4-cp314-cp314-macosx_10_15_x86_64.whl (322.6 kB view details)

Uploaded Jul 8, 2026 CPython 3.14macOS 10.15+ x86-64

sketchlog-1.2.4-cp313-cp313-win_amd64.whl (494.3 kB view details)

Uploaded Jul 8, 2026 CPython 3.13Windows x86-64

sketchlog-1.2.4-cp313-cp313-musllinux_1_2_x86_64.whl (1.4 MB view details)

Uploaded Jul 8, 2026 CPython 3.13musllinux: musl 1.2+ x86-64

sketchlog-1.2.4-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (340.9 kB view details)

Uploaded Jul 8, 2026 CPython 3.13manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

sketchlog-1.2.4-cp313-cp313-macosx_11_0_arm64.whl (314.9 kB view details)

Uploaded Jul 8, 2026 CPython 3.13macOS 11.0+ ARM64

sketchlog-1.2.4-cp313-cp313-macosx_10_13_x86_64.whl (322.8 kB view details)

Uploaded Jul 8, 2026 CPython 3.13macOS 10.13+ x86-64

sketchlog-1.2.4-cp312-cp312-win_amd64.whl (494.3 kB view details)

Uploaded Jul 8, 2026 CPython 3.12Windows x86-64

sketchlog-1.2.4-cp312-cp312-musllinux_1_2_x86_64.whl (1.4 MB view details)

Uploaded Jul 8, 2026 CPython 3.12musllinux: musl 1.2+ x86-64

sketchlog-1.2.4-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (340.8 kB view details)

Uploaded Jul 8, 2026 CPython 3.12manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

sketchlog-1.2.4-cp312-cp312-macosx_11_0_arm64.whl (314.8 kB view details)

Uploaded Jul 8, 2026 CPython 3.12macOS 11.0+ ARM64

sketchlog-1.2.4-cp312-cp312-macosx_10_13_x86_64.whl (322.7 kB view details)

Uploaded Jul 8, 2026 CPython 3.12macOS 10.13+ x86-64

sketchlog-1.2.4-cp311-cp311-win_amd64.whl (491.2 kB view details)

Uploaded Jul 8, 2026 CPython 3.11Windows x86-64

sketchlog-1.2.4-cp311-cp311-musllinux_1_2_x86_64.whl (1.4 MB view details)

Uploaded Jul 8, 2026 CPython 3.11musllinux: musl 1.2+ x86-64

sketchlog-1.2.4-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (337.9 kB view details)

Uploaded Jul 8, 2026 CPython 3.11manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

sketchlog-1.2.4-cp311-cp311-macosx_11_0_arm64.whl (312.7 kB view details)

Uploaded Jul 8, 2026 CPython 3.11macOS 11.0+ ARM64

sketchlog-1.2.4-cp311-cp311-macosx_10_9_x86_64.whl (319.9 kB view details)

Uploaded Jul 8, 2026 CPython 3.11macOS 10.9+ x86-64

sketchlog-1.2.4-cp310-cp310-win_amd64.whl (489.9 kB view details)

Uploaded Jul 8, 2026 CPython 3.10Windows x86-64

sketchlog-1.2.4-cp310-cp310-musllinux_1_2_x86_64.whl (1.4 MB view details)

Uploaded Jul 8, 2026 CPython 3.10musllinux: musl 1.2+ x86-64

sketchlog-1.2.4-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (336.5 kB view details)

Uploaded Jul 8, 2026 CPython 3.10manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

sketchlog-1.2.4-cp310-cp310-macosx_11_0_arm64.whl (311.2 kB view details)

Uploaded Jul 8, 2026 CPython 3.10macOS 11.0+ ARM64

sketchlog-1.2.4-cp310-cp310-macosx_10_9_x86_64.whl (318.3 kB view details)

Uploaded Jul 8, 2026 CPython 3.10macOS 10.9+ x86-64

File details

Details for the file sketchlog-1.2.4.tar.gz.

File metadata

Download URL: sketchlog-1.2.4.tar.gz
Upload date: Jul 8, 2026
Size: 267.7 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.1.0 CPython/3.13.13

File hashes

Hashes for sketchlog-1.2.4.tar.gz
Algorithm	Hash digest
SHA256	`e0da2fa0aed63ba498411dee9c35451fafb7426daac9b1bc9c92cd6c97d4397f`
MD5	`105c4a805c32e734aef0be500a2b0747`
BLAKE2b-256	`61ecd0a7320851bb1a4e8fee02db30a0f33f4d78d60516eea7abed47e9a40c29`

Algorithm	Hash digest
SHA256	`70bfd4d83be0e44113cc8a002f8b9f394220af73b7fb32e618b24357fbb8adc1`
MD5	`ce8ad702ec82db76c2f4a59f573f505b`
BLAKE2b-256	`a827ae7fd3158a975d43e9465c0bf896e3726722017906d14fe34756185d7946`

Algorithm	Hash digest
SHA256	`f30da3d6f2f53a376e3c06864b862d05da8e7adc8ce934d55333846c8b1dec51`
MD5	`b04bb323b69c1d5efbfe22684d70708e`
BLAKE2b-256	`c4c779c5233d659a490d2ba517c3cab8738450cc6da42a673535e0d2f5b3986d`

Algorithm	Hash digest
SHA256	`52f00e90a1c0109b908b00df171b0258d0de03f0a587f48d7200214fd191db4c`
MD5	`9fbb2ff355d1fa55f556daa98fb7a128`
BLAKE2b-256	`ab48f991455a30d23b1e7cd3355d0ceb17a72cd178445bafa3be7778d7856b7e`

Algorithm	Hash digest
SHA256	`768dbe52597468383d9b0396fbffe89a65c93f917203a2786a2e135f3296e202`
MD5	`2393f30d41c4fff60c1d15d2b2a61ab8`
BLAKE2b-256	`4d677e7715248508407af728010c280e32a8237c2a5b121a2dc55bf77f4a5a07`

Algorithm	Hash digest
SHA256	`6ceddf97b0d110ed81519ac810f8a43a1db839b3a795c21a6649a9a568582ec5`
MD5	`bfc84febf91562a8f34431e303c06991`
BLAKE2b-256	`9c923ceebc46db6a05b39aadf78b8a591f6799b5b0d7fa1b50f497010bd63559`

Algorithm	Hash digest
SHA256	`b8c145a59f20b67025c1713b15c11d0a2428daeeeca043bbdb6723e45b7dacfc`
MD5	`78e8d00fe0a9c5c0982d15c990d0d364`
BLAKE2b-256	`27ddbe93994ce3a1b3c0a7cc66ad607597bc64a781932bd535cbddd2c401d03e`

Algorithm	Hash digest
SHA256	`cfb955adb7a847399965fa4eb8df3bec6d58c7217a0c2ca737f411d061713ecc`
MD5	`ca654022ff55e053d4541b182239b54b`
BLAKE2b-256	`930a1c0272ae8082304bc0665b2417b3f79341b3242ec74f5686607556ab5a35`

Algorithm	Hash digest
SHA256	`221a036f19e138dbfd1fb992fd468aed0ce827d28139f47bdaa07ac255abd3cd`
MD5	`1fdc0d6bdc145fccb1b0f2d26629203b`
BLAKE2b-256	`dd22fa42b759696581c47cf5005891bfcd6a1714e87568e67e5a4daf07c47f2d`

Algorithm	Hash digest
SHA256	`dd5d95a4ffdd8d9b54b3f8400f98daaf15aa647fdb93d96afb7c1dc666f0e369`
MD5	`8855ed19ac8d499de679a4c31f47e8e6`
BLAKE2b-256	`a35e4ebba5d23192c68e3b5862181aff0ece8ed47ccf7fb4501f2b50afe35a62`

Algorithm	Hash digest
SHA256	`a0ca20512f51df0912d881e1051c50a7b0558c6f0f72f479772a69a9eb6f51bc`
MD5	`13021cc3a4c0439b73efd737968031c9`
BLAKE2b-256	`18267750e4ea184edcff1d9021a4a554a13c77395c1b781ffb21d2cffce570e3`

Algorithm	Hash digest
SHA256	`d5f5d7883b4de2d90b0d33884c1db8797be025e590c083e4567d684a16b8420b`
MD5	`633b4cfbe79e6c876617f98bcfa224b6`
BLAKE2b-256	`5a73cad52b1a78c03507af07ee27c70bf8b2f8402e20c440f67fb3cc87d1e904`

Algorithm	Hash digest
SHA256	`12e45af0ef314a4f45ee0365276f5793fe23b8baee111e57870ef07d2af68d3a`
MD5	`7b7529bbacea038d514de00bfe3c6901`
BLAKE2b-256	`0c95c0eb4ddf1dd8c6f51211336c9700bfba1cd4530084c8e8f87e37487a82c4`

Algorithm	Hash digest
SHA256	`cb6241eb43f3ad522c0827e7b7a657fd5ff774e2a4fc574384b1cb4fd774f026`
MD5	`6f6db1fafc78cdaa94fec3cdb4948610`
BLAKE2b-256	`bfdbb8f8f9d6506f0bbb121517c7168fb89032ccafa2fbd56c9768822a08dc64`

Algorithm	Hash digest
SHA256	`b18ae0e0aaf97e76ddaa46743846171f0f577770d05cf0ed65437fc8ea7caaa7`
MD5	`92948a24b83355548c8131ad469cc2c6`
BLAKE2b-256	`574d4b116059d9cb26d9bc6305a7d1291af897f555c603b0ddddd360844d416d`

Algorithm	Hash digest
SHA256	`c009ed4c7416f25fd5382602bdc8a37a917b4186fff7065938eb7b18f9e1c34a`
MD5	`051d4c9d081b540ae5d60e7d573b394e`
BLAKE2b-256	`56c0b065e778b327ed6807c6c3a38342cc400fb0022961cb150132d08d62172b`

Algorithm	Hash digest
SHA256	`4cfb4af5674f82538a3d8be12977019330ce5792aba67ac00bfd943fd8176b3e`
MD5	`6f58a0f3c0c53f8c417526e4af6b1022`
BLAKE2b-256	`bb3d141619c2d0e11801875b3acc7c397a27d0088289efd6175348e27833bd26`

Algorithm	Hash digest
SHA256	`edcb066df8f232a0ab42a3a20ba5decfca68ff00d350e8708320f6827ad66d3b`
MD5	`eae8daf4b0a342ee1d145bbb4eb804f5`
BLAKE2b-256	`d233400eb2ab9fa2b627056bc35b110ea0c4bea61438e21e4f8f672ea043c830`

Algorithm	Hash digest
SHA256	`e57aa7740a0e98476be05eec167cd1b169b1688745a55296842f27446ac5dbff`
MD5	`edcacbf50d17673657dfe61eac13e513`
BLAKE2b-256	`4136ea2f9fa4b9eaf8562193e30c348b8b99aa97c9f52d172f8d27873fae7466`

Algorithm	Hash digest
SHA256	`549a523034c56cef60302e1b26e3deb366bbc06ba7b8313584d340af148d92f0`
MD5	`5d84b0db19e34abdfaf7192b334341f2`
BLAKE2b-256	`e1027af87be2c2158e7c137c41ca3412862d019af0f7d590b7c980ac3eed6795`

Algorithm	Hash digest
SHA256	`43a8af0e577ebb92ecd183a55104b65aa6f8f8606407c47be72be9dade8cdc02`
MD5	`90739448f652ae04461ee513e5fc15cf`
BLAKE2b-256	`2b2de052091fa2a6b4436f2f0c26eea2c8200f436c8c3f905d8a10671861a0a7`

sketchlog 1.2.4

SketchLog

Why SketchLog

Live demo

Core capabilities

Architecture

Installation

Quickstart

Python embedded sketch

Python async client

TypeScript client

Go client

Deployment

Documentation

Development

What SketchLog is not

Community

License

Project links

Key dates

1 maintainer

Credits

GitHub Statistics

License expression

Requires

Provides Extra

Tags

Classifiers

Download files

Source Distribution

Built Distributions

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

Algorithm	Hash digest
SHA256	`581d3289a4dcbda65a433d19585dc526d6ec36d04354f90de28b8f8f451ba4b4`
MD5	`fd06703352e72851c4217749d9ca2825`
BLAKE2b-256	`513594005137e4d27380efa7cdceb9443137a5bfc78168c8d969ce552596e6fd`

Algorithm	Hash digest
SHA256	`33b34cf8ffc32a5dfbab26bf1adc8b18b9d2db07ae55d12b96bb45a7573761e6`
MD5	`416b9f8924f6bc593b4f41fbe75621ca`
BLAKE2b-256	`ead012e87283a7dc6721c1a7ba9d9f6e5e750a1a735a201a5d515b0ecd347e82`

Algorithm	Hash digest
SHA256	`0927c0addc3a2810620a4d9cc3d3de906b406094517a20bce661ff657cc84c5c`
MD5	`3176cfbb90d48efc4e7017195b017b7d`
BLAKE2b-256	`12a1ae47aa819d32f4ced56a2a10a283a8bc8fdfbb3f034fab628f8b7231d6af`

Algorithm	Hash digest
SHA256	`0f7312edd2a60abbcb2223b0fadbea981a857e069e8b05d21b5c5b3588c4cfcb`
MD5	`5759973e572cc0343274f0c62705edbe`
BLAKE2b-256	`cc01693a6ab96afca2c47cc2a3d2c3243340cf1307963c794435f0a5ed9fb993`

Algorithm	Hash digest
SHA256	`51b4ea0ac34991a1c67b6cfc94336cb92a8615429bf4e6bd088503e28d3cec49`
MD5	`a9a2ceae88711c25dc73483affc7ce6c`
BLAKE2b-256	`f1999c18cde98ea04ad1db3e6f0e893f0167e738d703d5badbec0f64999c9667`