# pinterest/secor

Secor is a service implementing Kafka log persistence

Repository: https://github.com/pinterest/secor
Canonical: https://ross.abutalabs.com/products/secor
Language: Java
License: Apache-2.0
License Family: permissive
Topics: kafka
Last push: 2026-03-10T08:39:10+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 71, release rhythm 35, longevity 100
- inputs: {"age_days": 4523, "days_push": 176, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1856, forks 537 (observed 2026-08-28T04:05:44.865281+00:00)

## What it is
Secor is a Java service that persists Apache Kafka log messages to object stores such as Amazon S3, Google Cloud Storage, Azure Blob Storage, and OpenStack Swift. It offers strong per-message consistency, fault tolerance, horizontal scalability, configurable size/time-based upload policies, and partitioned output paths for direct import into systems like Hive.

## Use cases
- persist kafka logs to s3
- archive kafka messages to cloud object storage
- export kafka topics to hive-partitioned files
- backup kafka data to a data lake
- move kafka events into s3 for offline analytics
- exactly-once kafka-to-object-storage pipeline
- run a scalable kafka consumer on kubernetes

## When to choose
- You need guaranteed, exactly-once persistence of Kafka messages into S3, GCS, Azure Blob Storage, or Swift
- You want time/partition-organized output (day/hour/minute) for direct Hive or data-lake imports
- You need a fault-tolerant, horizontally scalable Kafka consumer that can be scaled by simply adding or removing processes

## When to avoid
- You need real-time stream processing or per-message low-latency delivery rather than batched persistence to object storage
- Your target sink is not one of the supported object stores (S3, GCS, Azure Blob, Swift)
- You only need a simple single-node log dump without distributed consistency guarantees

## Facets
- artifact type: service
- maturity: stable
- function: streaming, etl
- domain: big-data, cloud-computing, analytics
- platform: jvm, cloud, self-hosted
- tags: kafka, apache-kafka, kafka-consumer, amazon-s3, google-cloud-storage, azure-blob-storage, openstack-swift, object-storage, log-persistence, log-archive, data-lake, hive, partitioned-output, exactly-once, java, data-engineering, messaging, kubernetes

## Member repositories
- pinterest/secor (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:44.865281+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:16:42.333602+00:00, confidence not recorded.
  - readme: https://github.com/pinterest/secor (fetched 2026-08-28T04:05:44.865281+00:00, sha 618f36924baf)
- Data as of 2026-08-30T08:39:29.467469+00:00.