Skip to content

Downloads

Public download links for the latest ELTMaestro packages. All links are direct (no login). The release bucket is not directory-browsable, so use the links on this page.

Current release line: 17.0.4 (engine / server). The -17.0.4 links below always point at the latest build of that line (refreshed on each publish). Refreshed 2026-07-20 from build 17.0.4.441 (master-claude-base). This build adds Snowflake query attribution — every Snowflake session the engine opens is tagged with the batch name, batch id and run number, so queries can be traced back to the batch that issued them via select parse_json(query_tag):batchName … against Snowflake's QUERY_HISTORY. No configuration needed; credentials are never included in the tag. It also carries the Salesforce SFSOURCE extractor (rolling-window delta, always-on deleted-record flag, and a Bulk API 2.0 fast path — see the step reference), on top of a server-side Job Metadata fix (editing a job's description/tags no longer leaves it locked to your session), the getVarData concurrency fix (no crash on concurrent SETVARIABLE), multipart S3 push (files >5 GB no longer fail with EntityTooLarge), the long-running-batch reaper watchdog, and the Snowflake closeOnCompletion guard — on top of the prior Jetty idle-timeout and Redshift measure-don't-estimate work — and ships the Windows client 2.5.2.20260722 (control tests now show on the Step Status tab — attached data-quality control tests, which run and write their results to the ABC database, are now listed on a workflow's Step Status tab alongside the step statuses, with expected/actual and pass/fail, so they're no longer invisible in the client; a Step Status name-alignment fix — the Step Status grid now shows each step's current name (matching the Runtime Status diagram) instead of the name recorded at the last run, so a renamed step no longer appears under both its old and new name; plus a canvas click-vs-drag fix — clicking a step on the editor canvas no longer nudges it out of place; a Set Job Metadata fix — the Config dialog now maps its fields correctly, so Description shows the job description instead of the job type and Apply no longer errors with value too long for type character varying(64); Apply now confirms, closes, and refreshes the workflow list; plus a Change History / Show Changes job version viewer with a read-only compare-mode canvas + per-step XML diff; Change Batch Cycle Type from the job right-click menu; a redesigned Manage Control Tests window — searchable grid with a separate editor; Redshift pipeline generators — bulk Export Jdbc To S3 Parquet and Import S3 Parquet To Redshift canvas wizards that build whole JDBCSOURCE→Parquet→S3 and COPY→table pipelines per selected table, with per-run S3-folder dedupe via step watermarks, optional first-column upsert, and a migratable $SYSTEM_DEFAULT_DATABASE target; a Read Uncommitted control for JDBC sources (per-step checkbox + a onstage_read_uncommitted System Setting — SQL Server extracts read at READ UNCOMMITTED); plus the prior 2.5.1 async JDBC test, Change Password dialog, canvas drag fix, and smooth scrolling).

Package Download Size md5 (current build) What it's for
Docker image https://eltmaestro.s3.amazonaws.com/docker/eltmaestro-17.0.4.tar.gz ~2.6 GB f1d2f84a20fb2861cddfe58562f101a2 Run the full ELTMaestro server as a container.
Linux installer (portable) https://eltmaestro.s3.amazonaws.com/docker/eltm-portable-17.0.4.tar.gz ~720 MB ca15f3c1f6ede42bda6865f83477c998 Bare-metal install on RHEL / Debian / Ubuntu.
Windows desktop client https://eltmaestro.s3.amazonaws.com/docker/ELTMAESTRO-2.5.3.20260724-portable-win-x64.zip ~80 MB 4ca897f9cf9cce3c36453ee28555a11c The desktop client used to connect to an ELTMaestro server. EV code-signed.

Checksums are for the current build (17.0.4.441 server / 2.5.3.20260724 client). The rolling -17.0.4 server objects change md5 on each publish; immutable per-build copies (eltmaestro-17.0.4.441.tar.gz, eltm-portable-17.0.4.441.tar.gz) are also on S3 with the same checksums.

What to do with each

  • Docker imagecurl -fLO <image-url> then docker load -i eltmaestro-17.0.4.tar.gz. Full walkthrough: Install from the Docker image.
  • Linux installercurl -fLO <installer-url>, extract, run install.sh. Full walkthrough: Install. The same tarball also installs an engine-only edge node — run install-engine.sh instead of install.sh to add a remote execution server that connects to the central audit DB: Edge node setup.
  • Windows desktop client — download the .zip, unzip, run ELTMAESTRO.exe, and point Server / Port at your ELTMaestro host (8181). See Getting started.

Versions

  • The -17.0.4 filenames track the current release line and are refreshed on each publish — use these for "the latest".
  • Specific pinned builds may also be present (e.g. eltmaestro-17.0.4.345.tar.gz) for reproducible installs. Substitute the exact build number in the filename if you need to pin.

Tip: the server (Docker image / Linux installer, version 17.0.4) and the Windows client (version 2.5.2) are versioned separately — that's expected.