LoginSign Up
← Back to all sparks
P

Parseable

ANALYTICS
Velocity5.0

Log analytics platform built for high-volume observability data

www.parseable.com ↗

Parseable is bolting real auth onto a log store — API keys, dataset permissions, Kafka IAM.

observabilitylog analyticsapi keysaccess controlkafka ingestionrust
◆Current state
The 2.7 through 2.9 line is dominated by authentication and access control. API keys arrived for ingestion and query, then as a managed feature, then had a security risk patched within weeks. Dataset-level user auth landed, OAuth sync was fixed, and the newest release adds AWS MSK IAM authentication over SASL/OAUTHBEARER plus a configurable OAuth provider for Kafka ingestion. Around it sit steady query and ingestion improvements: top-k in the counts API, insertion-time rather than data-time eviction, and field statistics reworked for high-volume ingestion.
◆Where it's heading
This is a project moving from single-tenant tool to something an organisation can hand to multiple teams: credentials that can be scoped and revoked, datasets that respect who is asking, and ingestion paths that authenticate against managed cloud services rather than static secrets. The speed with which an API key security risk appeared and was fixed shows the auth surface is new enough to still be settling.
◆Prediction
Expect the access control work to continue toward finer granularity — dataset permissions are in place, so per-key scoping and audit trails are the natural next steps. The Kafka OAuth provider being made configurable rather than MSK-specific suggests more managed-broker integrations follow.

◆Recent moves

  1. 15d ago

    Kafka ingestion gains AWS MSK IAM authentication

    Kafka ingestion can now authenticate to AWS MSK using IAM over SASL/OAUTHBEARER, with the OAuth provider made configurable rather than hardcoded to one vendor. Dataset-level user auth is fixed in the same release, continuing the access control thread that runs through this whole line.

    View source ↗
  2. 25d ago

    API key security risk patched weeks after launch

    A security risk in the API key implementation is fixed alongside an OAuth sync bug and a schema conflict — two weeks after API keys shipped as a feature. New authentication surfaces attract findings quickly, and this one did.

    View source ↗
  3. 1mo ago

    API keys land, plus top-k in the counts API

    API keys become a managed feature rather than a raw ingestion credential, and the counts API gains top-k within group by — a query shape that otherwise forces clients to over-fetch. The metastore also moves to a static form.

    View source ↗
  4. 1mo ago

    Eviction now tracks insertion time, not data time

    Retention switches to evicting by insertion time rather than the timestamp inside the data, which changes what actually gets deleted when logs arrive late or backfilled. Quiet, but the kind of semantic change an operator notices only after data disappears.

    View source ↗
  5. 1mo ago

    Field statistics tuned for high-volume ingestion

    A single performance change to field statistics under high ingestion volume, plus interface fixes in Prism. A point release with one commit of substance.

    View source ↗
  6. 1mo ago

    IRSA web identity for S3 and ingestion optimization

    S3 access begins loading IRSA web identity configuration, letting Parseable use Kubernetes service account identity instead of static keys, and the ingestion path is optimized. The container registry also moves to Quay, which is a packaging change users will notice on pull.

    View source ↗