Release Notes — 2.85

Covers the 2.85 release series. Most recent patch: 2.85.1.

Unless a subsection says otherwise, changes apply to both Monad SaaS and self-hosted.

New features & improvements

SaaS & self-hosted

• Component settings shown as a structured tree — the configuration panel on input, output, transform and enrichment detail pages now lays settings out as an expandable tree, with secrets and nested fields clearly marked. (2.85.1)
• Clearer Elasticsearch connection errors — when Test connection or pipeline startup can't reach the cluster, the error now includes the HTTP status and response, so bad credentials, a missing cluster privilege and a proxy block are easy to tell apart. (2.85.1)
• Least-privilege IAM policies for the SQS S3 inputs — the SQS S3, SQS S3 CloudTrail and SQS S3 GuardDuty pages now list only the four AWS actions the connectors call, with a copy-paste trust policy scoped to your queue and bucket. Docs → (2.85.1)
• Alert scope clarified — the alert docs now state that pipeline-scoped rules cover enabled pipelines only, and that a threshold rule treats a pipeline with no data in the window as 0, so it can catch a source that stopped sending. Docs → (2.85.1)
• Organization Logs contents defined — the Organization Logs input page now spells out what API Logs and Pipeline Logs contain. Docs → (2.85.1)
• Security guide: AWS access and buffered data — the security guide now explains how Monad assumes and handles your cross-account AWS role, and how records are buffered until they are delivered or purged. Docs → (2.85.1)
• Panther HTTP output: batching and size limits documented — covers Panther's 1 MB request limit and what a 413 response means. Docs → (2.85.1)

Self-hosted only

• AWS web-identity federation is a one-time hand-off — the AWS guide now shows that Monad needs your web-identity role ARN once, and that new target roles and AWS accounts are then yours to add without involving Monad. Docs → (2.85.1)

New connectors

None.

Fixes

• Elasticsearch output no longer loses records on failed bulk writes — write failures that were missed are now caught and retried, and an invalid index name is rejected when you save the output instead of failing on write. (2.85.1)
• CrowdStrike inputs with cloud autodiscover reach the right region — data requests now go to your tenant's regional API, and bad credentials are reported when the input starts. (2.85.1)
• Snyk Issues input no longer skips issues — issues updated within a second of the previous poll were missed; they are now collected. (2.85.1)
• Dialpad Change Log input receives webhook events — events sent to the input's webhook now reach the pipeline. (2.85.1)
• MCP server image patched — critical and high container vulnerabilities in the Monad MCP server image are resolved. (2.85.1)

Breaking changes

None.

Need help?

See the Monad docs for setup guides and reference, or reach out to Monad Customer Support at support@monad.com or via your dedicated Slack customer channel.